Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

10% Tumor Diameter Shrinkage on the First Follow-Up Computed Tomography Predicts Clinical Outcome in Patients With Advanced Renal Cell Carcinoma Treated With Angiogenesis Inhibitors: A Follow-Up Validation Study

10% Tumor Diameter Shrinkage on the First Follow-Up Computed Tomography Predicts Clinical Outcome... Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 Genitourinary Cancer: Renal, Bladder, and Testicular 10% Tumor Diameter Shrinkage on the First Follow-Up Computed Tomography Predicts Clinical Outcome in Patients With Advanced Renal Cell Carcinoma Treated With Angiogenesis Inhibitors: A Follow-Up Validation Study a,b c a,b d a,b a,b KATHERINE M. KRAJEWSKI, YOKO FRANCHETTI, MIZUKI NISHINO, ANDRE P. FAY, NIKHIL RAMAIYA, ANNICK D. VAN DEN ABBEELE, TONI K. CHOUEIRI a b Department of Imaging, Dana-Farber Cancer Institute, and Department of Radiology, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts, USA; Department of Biostatistics and Computational Biology, Dana-Farber Cancer Institute, and Department of Biostatistics, Harvard School of Public Health, Boston, Massachusetts, USA; Kidney Cancer Center, Dana-Farber Cancer Institute, Boston, Massachusetts, USA Disclosures of potential conflicts of interest may be found at the end of this article. Key Words. Renal cell carcinoma x VEGF-targeted therapy x RECIST 1.0 x Choi x Tumor shrinkage ABSTRACT Background. Vascular endothelial growth factor (VEGF)- stratified by response to the radiologic criteria. Receiver- targeted agents are standard therapies for metastatic renal cell operating characteristics (ROC) analysis yielded the optimal carcinoma (mRCC), associated with variable tumor shrinkage. threshold change in SLD, defining patients with prolonged Response Evaluation Criteria in Solid Tumors (RECIST) is of survival. limited utility in this setting, and other imaging changes are Results. More than 210%SLD significantly differentiated res- soughtto reliably predict outcome early.We aim to validate 10% ponders from nonresponders (median TTF 8.4 vs. 4.1 months, tumor shrinkage as the best early indicator of outcome. p5 .001), whereas partial response by RECIST did not (median Methods. In this institutional review board-approved, Health TTF6.9vs.5.5monthsinrespondersvs.nonresponders,p5.34). Insurance Portability and Accountability Act-compliant study, 210%SLD was also significantly predictive of OS (median 66 mRCC patients with 165 lesions on clinical trials of VEGF- OS 35.1 vs. 15.0 months in responders vs. nonresponders, targeted agents underwent thoracic and abdominal computed p 5 .003). ROC curve analysis yielded 29.3% in SLD as the tomography at baseline and at first follow-up after therapy. optimal threshold for response/no response. Measurements were performed according to RECIST and Conclusion. Ten percent tumor shrinkage is validated as a reli- tumor shrinkage of $10% decrease in sum of the longest able early predictor of outcome in mRCC patients receiving diameter (210%SLD). Correlation with time-to-treatment VEGF-targeted therapiesandmay provideapracticalmeasure to failure (TTF) and overall survival (OS) were compared and guide therapeutic decisions. The Oncologist 2014;19:507–514 Implications for Practice: Tumor shrinkage of 10% on first follow-up computed tomography after initiation of therapy predicts time-to-treatment failure and overall survival in this validation cohort of metastatic renal cell carcinoma patients treated with antiangiogenic therapies. Given the utilityof this response indicator in multiple reports, 10% tumor shrinkage may provide a simple and practical aid for therapeutic decision making if used in clinical trials and practice with a prospective evaluation. INTRODUCTION Molecular targeted therapies have revolutionized the treat- [2–6]. Many patients do not achieve 30% tumor shrinkage to ment of metastatic renal cell carcinoma (mRCC), resulting in indicate partial response in these reports, and it is commonly significant survival benefits in treated patients despite variable accepted that response evaluation criteria in solid tumors amounts of tumor shrinkage [1]. Antiangiogenic therapies are (RECIST) may be of limited utility in discriminating patients known to decrease tumor vascularization rather than result in responding to such therapies. direct cytotoxicity and have been associated with lesser Computed tomography (CT) is a robust, widely available degrees of tumor shrinkage than traditional antitumor agents, imaging modality routinely used in the assessment of patients’ with typical objective response rates ranging from 10% to 40% tumor burden throughout treatment and in monitoring Correspondence: Katherine M. Krajewski, M.D., Department of Imaging, Dana-Farber Cancer Institute, 450 Brookline Avenue, Boston, Massachusetts 02115, USA.Telephone: 617-582-8088; E-Mail: [email protected] Received October 10, 2013; accepted for publication February 21, 2014; first published online in The Oncologist Express on April 22, 2014. ©AlphaMed Press 1083-7159/2014/$20.00/0 http://dx.doi. org/10.1634/theoncologist.2013-0391 The Oncologist 2014;19:507–514 www.TheOncologist.com ©AlphaMed Press 2014 Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 508 10% Tumor Shrinkage and Outcome in Metastatic RCC patients for therapy-related toxicities. Several investigators patients had histologically confirmed stage IV mRCC. Patients have sought to optimize tumor response assessment via were treated at standard doses of study drug according to the imaging, such that imaging indicators better correlate with assigned protocol until they experienced disease progression, clinical benefit and survival outcomes early in the treatment severe or intolerable toxicity, or withdrew consent. Compliance course. Although the shortcomings of the conventional RECIST was checked after each cycle with a treatment diary. criteria in this settingareknown,there is no consensus asto the Imaging and Image Analysis preferred response assessment method. Several alternative Patients eligible for analysis included those with target lesions tumor shrinkage thresholds have been proposed, namely 10% by RECIST who underwent noncontrast-enhanced or contrast- and 20% long axis diameter reductions, either alone [7, 8] or in enhanced CTofthe chest, abdomen, and pelvisbefore andafter combination with other CT imaging response criteria, such as VEGF-targeted therapy initiation, with pre- and post-therapy attenuation [9–13]. Ten percent tumor shrinkage as an scans at the same institution.The routine oncology protocol was indicator of response has been supported in a study of patients used on multidetector CT scanners (64 detector row) using oral treated with sunitinib in a multicenter phase III trial [7] as contrast in all patients and intravenous contrast in patients with well as in another study of multiple imaging indicators of adequate estimated glomerular filtration rate (eGFR) and no response in patients treated with bevacizumab, sorafenib, known allergy to the contrast. Contrast-enhanced scans were and sunitinib [8]. Ten percent reduction in the long axis performed after 75–100 cc nonionic contrast administration diameter of target lesions is also a component of response (Ultravist; Bayer HealthCare, Leverkusen, Germany, http://www. accordingto Choi criteria [14], recently applied to patients with bayer.com;basedoneGFR),empiricallytimedwithchestimages mRCC [9, 12, 13]. Twenty percent long axis reductions are obtained in the arterial phase (30-second delay) and abdominal components of favorable response in the recently developed images obtained in the portal venous phase (70-second delay). size and attenuation CT (SACT) and morphology, attenuation, Images were reviewed and measured on a Picture Archiving and size, and structure (MASS) criteria [10, 11]. Communication System (Centricity; General Electric, Milwaukee, Inapreviousstudyof70mRCCpatientstreatedwithfirst-line WI, http://www.gehealthcare.com). sunitinib, sorafenib, or bevacizumab, four criteria of early Up to three target lesions in each patient were selected on postimaging changes were compared, including RECIST 1.0, 10% the baseline CT, according to RECIST 1.0 [20], by a single board- tumor diameter shrinkage, CT attenuation of tumor, and Choi certified radiologist with 8 years of experience in cancer criteria in assessing time-to-treatment failure (TTF) and overall imaging, blinded to patient outcomes. At baseline and first survival (OS) [8]. Ten percent tumor shrinkage at first follow-up follow-up, the longest axial axis of each target was recorded to was demonstrated statistically significant in predicting TTF and the nearest millimeter, as described previously [8]. Sum of the OS, whereas other criteria were not. However, given the longest diameter (SLD) of the targets was recorded at baseline retrospective nature of the study, it is necessary to validate the and at first follow-up of all patients. results in an independent cohort and reproduce the association RECIST response, including partial response (PR) and between 10% tumor shrinkage and survival, to propose use in progressive disease (PD), was recorded at CT follow-up, based future prospective clinical trials and clinical practice. on$30% decrease in SLD for PR and$20% increase in SLD for The purpose of the present study is to determine whether PD [20]. Patients not meeting either were tabulated as stable 10% tumor shrinkage can differentiate responders from disease (SD).Whereas RECISTresponse accordingto RECIST 1.0 nonresponders in an independent validation cohort of mRCC was studied in this report, the RECIST categorizations are patients enrolled in prospective clinical trials of vascular thought to be comparable to those according to the updated endothelial growth factor (VEGF) inhibitors. We hypothesize RECIST 1.1 criteria, because the definitions of PR, PD, and SD that10%tumorshrinkageatfirstfollow-up,asdemonstratedin are unchanged [21, 22]. Ten percent decrease in SLD was also our previous report and others [7, 8], represents a more examined as an independent predictor of outcome. clinically meaningful response assessment than RECIST 1.0. If Clinical outcome data were gathered by review of medical found reproducible, a 10% tumor shrinkage criterion could be records, including date of treatment failure (progression as widely applicable to mRCC patients, to identify patients likely determined by RECIST) and date of death. TTF was defined as to benefit from treatment early after initiation of therapy. time from treatment initiation until failure, including pro- gression and death. Patients who did not progress were PATIENTS AND METHODS censored on the date of last follow-up. OS was defined as the time from treatment initiation until death. Alive patients were Patients and Treatment censored on the date of last follow-up. The study sample consists of patients with metastatic renal cell carcinoma enrolled in six recent and ongoing multicenter, open- Statistical Analysis label studies of VEGF-targeted agents (vatalanib, sunitinib, Patients were classified into responders (PR) versus SD or PD, sorafenib,pazopanib,foretinib,andtivozanib)(3,15–19).Patients according to RECIST 1.0.The survival distribution for patients in were part of approved Institutional Review Board (IRB) pro- the two groups was estimated according to the Kaplan-Meier tocols for advanced renal cell carcinoma (RCC) at the institution method [23]. The difference in the survival distribution in which baseline and follow-up clinical data were prospectively between the groups was evaluated using the log-rank test. collected, and this study was approved by the IRB and is Health An analysis of RECIST 1.0 clinical benefit, PR or SD versus PD, Insurance Portability and Accountability Act compliant. Patients was not performed as only one patient had PD at the time of enrolled in these trials who had been included in a previous first follow-up, rendering this invalid. To explore any effect study cohort were excluded from this validation cohort [8]. All modifier among baseline characteristics in the relationships The ©AlphaMed Press 2014 Oncologist Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 Krajewski, Franchetti, Nishino et al. 509 Table 1. Patient demographic and disease characteristics of cohort had clear cell RCC histology (59%), 22 patients had papillary (33%), 2 patients each had chromophobe and the 66 patients unavailable histologic subtype, and 1 patient had predominant Characteristics Frequency Percent sarcomatoid features in addition to having clear cell histology. Age (median, range) 60.9 (32–84) The median age at targeted therapy initiation was 60.9 Gender years.Twenty-one patients (32%) were treated with vatalanib, 17 patients (26%) with foretinib, 10 patients (15%) each with Male 46 70 pazopanib and sunitinib, 6 patients (9%) with sorafenib, and 2 Female 20 30 patients (3%) with tivozanib. Median time from baseline CT to Histology drug initiation was 13 days, and median time from baseline to Clear cell 39 59 first post-treatment CT was 62 days (range 20–172 days). The Papillary 22 33 observed outlier of 172 days in time from baseline to first post- Chromophobe 2 3.0 treatment CT lay in 7 times the interquartile range (52 to 67 Sarcomatoid with clear cell 1 1.6 days) above the third quartile among all samples. All 66 Not specified 2 3.0 patients were evaluated according to RECIST 1.0 and 10% tumor shrinkage thresholds. Medication Using the selected target lesions for this study, most Vatalanib 21 31.8 patients (59 patients, 89%) were SD according to RECIST 1.0 at Foretinib 17 25.8 first post-treatment CT, whereas 6 were PR (9%) and 1 was PD Pazopanib 10 15.2 (2%). Using 10% tumor shrinkage to indicate response, 33 Sorafenib 6 9.1 patients achieved 10% tumor shrinkage and were considered Sunitinib 10 15.2 responders (50%; Fig. 1), whereas 33 patients achieved less Tivozanib 2 3.0 than 10% tumor shrinkage, deemed nonresponders. Pro- Reason for off study portionalchangesofSLD inreferencetobaselineinall but one of PD 38 57.6 the study patients at the time of first follow-up CT ranged from 52% decrease in SLD to 20% increase in the SLD of target lesions On trial 2 3.0 (Fig. 2). One patient with a single measurable target lesion in the Toxicity 24 36.4 peritoneum atbaselineexperienced significanttumorshrinkage Withdrew consent 2 3.0 such that the target lesion was no longer measurable at first CT size change from baseline 210.8 (2100, 20) follow-upCT,although nontargetascitesfluidremained present, to first follow-up CT (%) rendering the patient a PR according to RECIST. (median, range) No patient who achieved 10% tumor shrinkage at first Interval from baseline 13 [3, 27] CT to treatment initiation follow-up on retrospective assessment developed a new lesion (days) (median, range) or was deemed as having progressive disease at the time of the Interval from baseline CT to 62 (20, 172) original response assessment. post-treatment CT (days) (median, range) Association Between Early Post-Therapy Imaging Follow-up from registration 22.9 (1.3, 94.4) Changes and Clinical Outcome to the last documentation The TTF for patients with PR (median 6.9 months) was not or death (months) (median, range) significantly different from patients with SD or PD (median 5.5 Abbreviations: CT, computed tomography; PD, progressive disease. months, p5 .34; Fig. 3A). Similarly, the OS for patients with PR (median not reached, NR) was not significantly different from between decrease in SLD and survival outcomes, bivariate patients with SD or PD (median 25.5 months, p5 .72; Fig. 3B). analyses were performed using Cox proportional hazard Tumor shrinkage of $10% decrease in SLD significantly models in an exploratory fashion. differentiated responders (n 5 33, median TTF 8.4 months) Receiver-operating characteristics (ROC) analysis was used to from nonresponders (n 5 33, median TTF 4.1 months, p 5 determine the optimal threshold change in SLD at first CT follow- .001;Fig.4A).Tenpercentdecreasein SLDwasalsosignificantly uptodefinepatientswithsurvivalendpoint(deathorcensoredat predictive of overall survival (with responders OS of 35.1 the end of the study), similar to Thiam et al. [7]. After confirming months vs. nonresponders OS of 15.0 months, p 5 .003; validity of the area under the ROC curve (AUC) in an accuracy test Fig. 4B). Bivariate Cox proportional hazards models showed using a traditional point-system classification, Youden’sindex in that 10% decrease in SLD was a significantly robust predictor of the ROC analysis was calculated for the series of Ki-67 values TTF (p5 .001) and OS (p5 .01), respectively, after controlling observed in the data.We identified the threshold that minimized for either baseline age, gender, drug, histology, or Memorial the false-positive and false-negative rates for best detecting Sloan-Kettering Cancer Center risk (low/intermediate vs. prolonged survival, selecting patients with treatment benefit. high), whereas none of the controlled factors was significant. RESULTS Threshold Evaluation by ROC Analysis Patients For tumor shrinkage threshold evaluation, ROC curve analysis Sixty-six patients with 165 target lesions were eligible for the yielded29.3%inSLD(maximumYoudenindexscore0.28)asthe study and evaluated (Table 1). Thirty-nine patients of the optimal threshold for response/no response with respect to OS. www.TheOncologist.com ©AlphaMed Press 2014 Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 510 10% Tumor Shrinkage and Outcome in Metastatic RCC Figure 1. A 73-year-old man with metastatic renal cell carcinoma treated with foretinib. Axial contrast-enhanced computed tomography (CT) images at baseline (A and B) demonstrate target right adrenal and retroperitoneal metastases (black measurement lines), measuring 31 mm and 58 mm in long axis, respectively. Axial contrast-enhanced CT at first follow-up after treatment initiation (C and D) demonstrates approximately 10% decrease in the sum of the longest diameter of the targets (black measurement lines), measuring 28 mm and 52 mm, respectively. survival in 334 treated patients, including ROC analysis of DISCUSSION various other shrinkage thresholds [7]. The advocated 210% We evaluated RECISTresponse and tumor shrinkage as practical threshold in this study and the 230% threshold for PR and reproducible imaging predictors of benefit from various according to RECIST were reached after the first cycle of VEGF-targeted therapies in mRCC patients treated in recent and therapy in 73% and 19% of cases, respectively [7].Ten percent ongoingclinicaltrials.Inourstudyof66patients,asizablecohort tumor shrinkage was also a significant predictor of TTF and OS, with respect to similar literature, a 10% decrease in target SLD whereas RECIST and Choi criteria were not predictive in was a reliable predictor of TTF and OS outcomes on first follow- another study of patients treated with bevacizumab, sunitinib, up CT. This finding lends further support to the 10% tumor and sorafenib [8]. The current study lends further support to shrinkage threshold as an indicatorof response in this setting, as the applicability and utility of 10% tumor shrinkage as an early previously advocated in the literature [7, 8]. Such early changes indicatorof response in mRCC patients treatedwith avariety of in post-therapy imaging enable us to distinguish responder VEGF-targeted therapies, as class-specific tumor response patients and aid in patient management in three ways: (a) criteria rather than drug-specific. permitting patients who achieve this degree of tumor shrinkage Choi criteria have been applied to mRCC patients treated to continue treatment with greater confidence, (b) limiting with VEGF-targeted therapies, with differing conclusions on unnecessary toxicities to patients without evidence of treat- the applicability of these criteria in different reports. Choi ment benefit, and (c) enabling these patients to consider other criteria similarly incorporate 10% tumor shrinkage into “re- therapies. sponse,”or 15% decrease in mean tumor CT attenuation. In one The 10% tumor shrinkage threshold as an indicator of such report by van der Veldt et al. [9], Choi criteria better response to VEGF-targeted therapies in mRCC patients has delineated responders and nonresponders to therapy compared been supported by analysis ofa prior phase III studyofsunitinib with RECIST, but did not indicate early progression. Thereafter, in which 10% tumor shrinkage was the best predictor of Choicriteria were foundto earlydiscriminate patients benefiting The ©AlphaMed Press 2014 Oncologist Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 Krajewski, Franchetti, Nishino et al. 511 The unmet clinical need and utility of more generalized tumor response criteria are emphasized by various criteria evaluated in recent literature. A tumor response criterion that is generalizable as well as indicative of treatment benefit, such as one based on a tumor shrinkage threshold, is preferred and sought in this population. CT scanning is widely available and serves multiple purposes in the care of patients with widespread metastatic renal cell carcinoma. It is commonly used to assess the following: (a) patients’overall tumor burden intermsofgrowthpatterns anddistribution; (b)individual sites of metastatic disease to anticipate and limit troublesome disease-related complications via alternative, focused treat- ments (such as surgery or radiation to spinal metastases that may result in cord compression); and (c) evidence of treatment-related toxicities, potentially before the patient develops symptoms (such as pneumatosis intestinalis and/or hemorrhage in VEGF-targeted treatments) (25, 26). A simple, reliablemeasurementschemathathasabettercorrelationwith survival than RECIST can easily be incorporated into routine clinical practice at any center, regardless of scan protocol. Furthermore, single kidney patients with elevated eGFR or those with allergy may be assessed accordingly using the 10% Figure 2. Range of change in the sum long axis diameter of target tumorshrinkage, withoutthe necessityofintravenouscontrast. lesions at first follow-up computed tomography (CT). Responder patients with 10% tumor shrinkage at the time of the first follow- There are several limitations to this study. For instance, up CT (blue bars) had median time-to-treatment failure (TTF) and the imaging review was performed retrospectively, by a single overall survival (OS) of 8.4 and 35.1 months, respectively, whereas oncoradiologist, and patients were treated at a single in- nonresponder patients who did not achieve at least 10% tumor stitution. Given the small number of patients and the shrinkage (red bars) had median TTF and OS of 4.1 and 15.0 retrospective nature of the study, comparison of the 10% tumor months (p 5 .0014 and .0032, respectively). shrinkage threshold with the RECIST criteria is limited and should not be overinterpreted. In fact, in this study, we applied from treatment in at least one other report [12], although they the RECIST definitions of partial response, stable disease, and didnotsignificantlycorrelatewithsurvivaloutcomesinothers[8, progressive disease to the time of first follow-up CT rather than 13, 24]. In each of these studies, Choi criteria were more likely to assess RECIST response at all study time points. Of note, most characterizepatientsasrespondersthanothercriteriaingeneral, RECISTresponses do happen within 2 months (60 days median). regardless of outcome. Technical CT parameters, contrast, CT According to Thiam et al. [7], 64% of partial responses were attenuation measurementvariability, and patientvariables most achievedwithintwocyclesofsunitinibtreatment.Therefore,itis likely contribute to the differences in results observed. In likely that we captured most patients who achieved partial addition, no singular attenuation change threshold has been response forthe best response duringtheirtreatmentcourse,as found to correlate with survival [8], raising questions regarding a result of the timing of the first follow-up CTs. theapplicabilityofthe15%decreasedattenuationcomponentof Our purpose was to evaluate early changes in tumor burden this response schema in the mRCC population. on first CT follow-up and correlate these imaging changes with Similarly, the SACT and subsequent MASS criteria were survival. Although nadir tumor burden or “best overall re- developed to delineate response in mRCC patients treated with sponse” is important, this information is not available until sunitinib and sorafenib. These criteria include 20% tumor a patient has finished therapy. Early tumor shrinkage is useful shrinkage, marked central necrosis, or marked decreased because it can be assessed prospectively in all patients. Our attenuation as components of favorable response, whereas results show that early tumor shrinkage is informative, although 20% increase in size, new lesions, marked central fill-in, or new the results of our study should be interpreted in the context of enhancement are features of unfavorable response [10, 11]. the study design. Another persistent limitation in this validation PretherapyschemahavebeenrecentlycombinedwiththeMASS cohort is that it may be difficult to place patients close to the criteria to improve accuracy [24], offering a unique clinical and threshold (9% decrease) in the appropriate response category, imaging response model. However, drawbacks of these criteria but this would be true for any criteria using a threshold. includearbitrarycorrelationwithprogression-freesurvival.250 Measurement reproducibility in this cohort was separately days, exclusion of lung lesions, and volumetric measurements evaluated [27]. In this study, we found 10% tumor shrinkage was necessitating another workstation in SACT. In contrast to these a reproducible radiologic response indicator when baseline and methods, the 10% tumor shrinkage assessment method de- follow-up studies were measured by a single radiologist. scribed in this work may be advantageous because of the However, 10% tumor shrinkage was within the range of simplicity of single long axis diameter measurements of only interobserver measurement variability. The validation study three target lesions, which can be performed on a routineclinical findings must be considered in this context.Some of the benefits basiswithoutaddedtimeoraseparateadditionalworkstation,as of the patient cohort in this study include fairly homogeneous is needed in volumetric measurements. timing of the baseline and follow-up scans. Patients in this study www.TheOncologist.com ©AlphaMed Press 2014 Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 512 10% Tumor Shrinkage and Outcome in Metastatic RCC Figure 3. Time to treatment failure (TTF) and overall survival (OS) stratified by RECIST response at first follow-up computed tomography. (A): TTF stratified by RECIST response (PR) versus no response (SD 1 PD). TTF was not significantly different between RECIST responders and nonresponders (6.9 vs. 5.5 months, p 5 .3418). Partial response occurred in 6 patients at first follow-up, 6 events; 60 patients were nonresponders (SD 1 PD; 57 events and 3 censored). (B): OS stratified by RECIST response (PR) versus no response (SD 1 PD). OS was not significantly different between RECIST responders and nonresponders (not reached vs. 25.5 months, p5 .7185). Partial response occurred in 6 patients at first follow-up, with 3 patients deceased and 3 censored; 60 patients were nonresponders (SD1 PD; 43 deceased and 17 censored). Abbreviations: PD, progressive disease; PR, partial response; RECIST, Response Evaluation Criteria In Solid Tumors; SD, stable disease. The ©AlphaMed Press 2014 Oncologist Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 Krajewski, Franchetti, Nishino et al. 513 Figure 4. Time-to-treatment failure (TTF) and overall survival (OS) stratified by first follow-up computed tomography response according to .10%decreaseinsizeversusnoresponse.(A):TTFstratifiedbyresponseaccordingto$10%decreaseinsizeversusnoresponse(,10%decrease in size). TTF was significantly different between responders who achieved 10% decrease in size and nonresponders who did not (8.4 vs. 4.1 months, p 5 .0014). The $10% tumor shrinkage occurred in 33 patients at first follow-up, with 30 events and 3 censored; 33 patients were nonresponders(33events).(B):OSstratifiedbyresponseaccordingto$10%decreaseinsizeversusnoresponse(,10%decreaseinsize).OSwas significantlydifferent betweenresponders whoachieved10% decrease insizeand nonresponders whodidnot(35.1vs. 15.0months, p5.0032). The $10% tumor shrinkage occurred in 33 patients at first follow-up, with 18 patients deceased and 15 censored; 33 patients were nonresponders, with 28 patients deceased and 5 censored. Abbreviation: SLD, sum of the longest diameter. www.TheOncologist.com ©AlphaMed Press 2014 Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 514 10% Tumor Shrinkage and Outcome in Metastatic RCC were on clinical trials of various VEGF-targeted agents, whereas AUTHOR CONTRIBUTIONS patientsinapreviousstudywerenottreatedontrials[8].Second, Conception/Design: Katherine M. Krajewski, Nikhil Ramaiya, Toni K. Choueiri data and conclusions from this study may be applicable to mRCC Provision of study material or patients: Toni K. Choueiri Collection and/or assembly of data: Katherine M. Krajewski, Mizuki Nishino, patients treated with VEGF-targeted therapies in general. Andre´ P. Fay, Toni K. Choueiri Data analysis and interpretation: Katherine M. Krajewski, Yoko Franchetti, CONCLUSION Mizuki Nishino, Andre´ P. Fay, Nikhil Ramaiya, Annick D.Van den Abbeele,Toni K. Choueiri The 10% tumor shrinkage threshold is predictive of survival in Manuscript writing: Katherine M. Krajewski, Yoko Franchetti, Mizuki Nishino, this validation cohort of mRCC patients treated with VEGF- Nikhil Ramaiya, Annick D. Van den Abbeele, Toni K. Choueiri Final approval of manuscript: Katherine M. Krajewski, Yoko Franchetti, Mizuki targeted therapies. ROC analysis identified 9.3% decrease in SLD Nishino, Andre P. Fay, Nikhil Ramaiya, Annick D. Van den Abbeele at first follow-up as the best size predictor of time-to-treatment failure.Thus, a reduction of 10% in the sum longaxis diameter of DISCLOSURES targets is validated as a reliable, early predictor of outcome in Katherine M. Krajewski: General Electric (RF); Toni K. Choueiri: Pfizer, this setting. Given the utility of this response indicator in Aveo, Exilixis, Novartis (C/A).The other authors indicated no financial multiple reports,the 10% tumorshrinkage may provide a simple relationships. (C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert and practical aid for therapeutic decision making if used in testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/ clinical trials and practice with a prospective evaluation. inventor/patent holder; (SAB) Scientific advisory board REFERENCES 1. Heng DY, Xie W, Regan MM et al. External Importance of size and attenuation on contrast- 19. Hutson TE, Rathmell K, Hudes GR et al. A phase validation and comparison with other models of the enhanced CT. AJR Am J Roentgenol 2010;194: II biomarker assessment of tivozanib in oncology International Metastatic Renal Cell Carcinoma 157–165. (BATON) trial in patients (pts) with advanced renal Database Consortium prognostic model: A popula- cell carcinoma (RCC). J Clin Oncol 2012;30(suppl): 11. Smith AD, Shah SN, Rini BI et al. Morphology, tion based study. Lancet Oncol 2013;14:141–148. abstract TPS4686. attenuation, size, and structure (MASS) criteria: 2. Yang JC, Haworth L, Sherry RM et al. A Assessing response and predicting clinical outcome 20. Therasse P, Arbuck SG, Eisenhauer EA et al. randomized trial of bevacizumab, an anti-vascular in metastatic renal cell carcinoma on antiangiogenic New guidelines to evaluate the response to endothelial growth factor antibody, for metastatic targeted therapy. AJR Am J Roentgenol 2010;194: treatment in solid tumors. European Organization 1470–1478. for Research and Treatment of Cancer, National renal cancer. N Engl J Med 2003;349:427–434. Cancer Institute of the United States, National 12. Schmidt N, Hess V, Zumbrunn T et al. Choi 3. Motzer RJ, Michaelson MD, Redman BG et al. Cancer Institute of Canada. J Natl Cancer Inst 2000; response criteria for prediction of survival in Activity of SU11248, a multitargeted inhibitor of 92:205–216. patients with metastatic renal cell carcinoma vascular endothelial growth factor receptor and treated with anti-angiogenic therapies. Eur Radiol 21. EisenhauerEA,TherasseP,BogaertsJ etal.New platelet-derived growth factor receptor, in patients 2013;23:632–639. response evaluation criteria in solid tumours: with metastatic renal cell carcinoma. J Clin Oncol Revised RECIST guideline (version 1.1). Eur J Cancer 2006;24:16–24. 13. Hittinger M, Staehler M, Schramm N et al. 2009;45:228–247. Course of size and density of metastatic renal cell 4. Motzer RJ, Hutson TE, Tomczak P et al. Sunitinib carcinoma lesions in the early follow-up of molec- 22. Nishino M, Jagannathan JP, Ramaiya NH et al. versus interferon alfa in metastatic renal-cell ular targeted therapy. Urol Oncol 2012;30:695–703. Revised RECIST guideline version 1.1: What oncol- carcinoma. N Engl J Med 2007;356:115–124. ogists want to know and what radiologists need to 14. Choi H, Charnsangavej C, Faria SC et al. 5. Escudier B, Eisen T, Stadler WM et al; TARGET know. AJR Am J Roentgenol 2010;195:281–289. Correlation of computed tomography and positron Study Group. Sorafenib in advanced clear-cell renal- emission tomography in patients with metastatic 23. Kaplan EL, Meier P. Nonparametric estimation cell carcinoma. N Engl J Med 2007;356:125–134. gastrointestinal stromal tumor treated at a single from incomplete observations. J Am Stat Assoc 6. Motzer RJ, Michaelson MD, Rosenberg J et al. institution with imatinib mesylate: Proposal of new 1958;53:457–481. Sunitinib efficacy against advanced renal cell computed tomography response criteria. J Clin 24. Smith AD, Shah SN, Rini BI et al. Utilizing pre- carcinoma. J Urol 2007;178:1883–1887. Oncol 2007;25:1753–1759. therapy clinical schema and initial CT changes to 7. Thiam R, Fournier LS, Trinquart L et al. Optimiz- 15. George D, Michaelson D, Oh WK et al. Phase I predict progression-free survival in patients with ing the size variation threshold for the CT evaluation study of PTK787/ZK 222584 (PTK/ZK) in metastatic metastatic renal cell carcinoma on VEGF-targeted of response in metastatic renal cell carcinoma renal cell carcinoma. Proc Am Soc Clin Oncol 2003; therapy: A preliminary analysis. Urol Oncol 2013;31: treated with sunitinib. Ann Oncol 2010;21:936–941. 22: abstract 1548. 1283–1291. 8. Krajewski KM, Guo M,Van den Abbeele AD et al. 16. Stadler WM, Figlin RA, McDermott DF et al; 25. Coriat R, Ropert S, Mir O et al. Pneumatosis Comparison of four early posttherapy imaging ARCCS Study Investigators. Safety and efficacy intestinalis associated with treatment of cancer changes (EPTIC; RECIST 1.0, tumor shrinkage, results of the advanced renal cell carcinoma patients with the vascular growth factor receptor computed tomography tumor density, Choi criteria) sorafenib expanded access program in North tyrosine kinase inhibitors sorafenib and sunitinib. in assessing outcome to vascular endothelial growth America. Cancer 2010;116:1272–1280. Invest New Drugs 2011;29:1090–1093. factor-targeted therapy in patients with advanced 17. MotzerRJ,Hutson TE,ReevesJetal.Randomized 26. Sonpavde G, Bellmunt J, Schutz F et al. The renal cell carcinoma. Eur Urol 2011;59:856–862. open-label phase III trial of pazopanib versus sunitinib double edged sword of bleeding and clotting from 9. van der Veldt AA, Meijerink MR, van den in first-line treatment of patients withmetastatic renal VEGF inhibition in renal cancer patients. Curr Oncol Eertwegh AJ et al. Choi response criteria for early cell carcinoma (MRCC): Results of the COMPARZ trial. Rep 2012;14:295–306. prediction of clinical outcome in patients with Paper presented at: 2012 ESMO Congress; October 1, 27. Krajewski KM, Nishino M, Franchetti Y et al. metastatic renal cell cancer treated with sunitinib. 2012, Abstract LBA8; Vienna, Austria. Intraobserver and interobserver variability in com- Br J Cancer 2010;102:803–809. 18. Choueiri TK, Vaishampayan U, Rosenberg JE puted tomography size and attenuation measure- 10. Smith AD, Lieber ML, Shah SN. Assessing tumor et al. Phase II and biomarker study of the dual MET/ ments in patients with renalcell carcinomareceiving response and detecting recurrence in metastatic VEGFR2 inhibitor foretinib in patients with papillary antiangiogenic therapy: Implications for alternative renal cell carcinoma on targeted therapy: renal cell carcinoma. J Clin Oncol 2013;31:181–186. response criteria. Cancer 2014;120:711–721. EDITOR’S NOTE: See the related commentary by Helen X. Chen et al. on page 439. The ©AlphaMed Press 2014 Oncologist http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The Oncologist Oxford University Press

10% Tumor Diameter Shrinkage on the First Follow-Up Computed Tomography Predicts Clinical Outcome in Patients With Advanced Renal Cell Carcinoma Treated With Angiogenesis Inhibitors: A Follow-Up Validation Study

Download PDF
8 pages

Loading next page...
 
/lp/oxford-university-press/10-tumor-diameter-shrinkage-on-the-first-follow-up-computed-tomography-Et8OYaD1ZK

References (29)

Publisher
Oxford University Press
Copyright
Copyright © 2022 Oxford University Press
ISSN
1083-7159
eISSN
1549-490X
DOI
10.1634/theoncologist.2013-0391
Publisher site
See Article on Publisher Site

Abstract

Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 Genitourinary Cancer: Renal, Bladder, and Testicular 10% Tumor Diameter Shrinkage on the First Follow-Up Computed Tomography Predicts Clinical Outcome in Patients With Advanced Renal Cell Carcinoma Treated With Angiogenesis Inhibitors: A Follow-Up Validation Study a,b c a,b d a,b a,b KATHERINE M. KRAJEWSKI, YOKO FRANCHETTI, MIZUKI NISHINO, ANDRE P. FAY, NIKHIL RAMAIYA, ANNICK D. VAN DEN ABBEELE, TONI K. CHOUEIRI a b Department of Imaging, Dana-Farber Cancer Institute, and Department of Radiology, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts, USA; Department of Biostatistics and Computational Biology, Dana-Farber Cancer Institute, and Department of Biostatistics, Harvard School of Public Health, Boston, Massachusetts, USA; Kidney Cancer Center, Dana-Farber Cancer Institute, Boston, Massachusetts, USA Disclosures of potential conflicts of interest may be found at the end of this article. Key Words. Renal cell carcinoma x VEGF-targeted therapy x RECIST 1.0 x Choi x Tumor shrinkage ABSTRACT Background. Vascular endothelial growth factor (VEGF)- stratified by response to the radiologic criteria. Receiver- targeted agents are standard therapies for metastatic renal cell operating characteristics (ROC) analysis yielded the optimal carcinoma (mRCC), associated with variable tumor shrinkage. threshold change in SLD, defining patients with prolonged Response Evaluation Criteria in Solid Tumors (RECIST) is of survival. limited utility in this setting, and other imaging changes are Results. More than 210%SLD significantly differentiated res- soughtto reliably predict outcome early.We aim to validate 10% ponders from nonresponders (median TTF 8.4 vs. 4.1 months, tumor shrinkage as the best early indicator of outcome. p5 .001), whereas partial response by RECIST did not (median Methods. In this institutional review board-approved, Health TTF6.9vs.5.5monthsinrespondersvs.nonresponders,p5.34). Insurance Portability and Accountability Act-compliant study, 210%SLD was also significantly predictive of OS (median 66 mRCC patients with 165 lesions on clinical trials of VEGF- OS 35.1 vs. 15.0 months in responders vs. nonresponders, targeted agents underwent thoracic and abdominal computed p 5 .003). ROC curve analysis yielded 29.3% in SLD as the tomography at baseline and at first follow-up after therapy. optimal threshold for response/no response. Measurements were performed according to RECIST and Conclusion. Ten percent tumor shrinkage is validated as a reli- tumor shrinkage of $10% decrease in sum of the longest able early predictor of outcome in mRCC patients receiving diameter (210%SLD). Correlation with time-to-treatment VEGF-targeted therapiesandmay provideapracticalmeasure to failure (TTF) and overall survival (OS) were compared and guide therapeutic decisions. The Oncologist 2014;19:507–514 Implications for Practice: Tumor shrinkage of 10% on first follow-up computed tomography after initiation of therapy predicts time-to-treatment failure and overall survival in this validation cohort of metastatic renal cell carcinoma patients treated with antiangiogenic therapies. Given the utilityof this response indicator in multiple reports, 10% tumor shrinkage may provide a simple and practical aid for therapeutic decision making if used in clinical trials and practice with a prospective evaluation. INTRODUCTION Molecular targeted therapies have revolutionized the treat- [2–6]. Many patients do not achieve 30% tumor shrinkage to ment of metastatic renal cell carcinoma (mRCC), resulting in indicate partial response in these reports, and it is commonly significant survival benefits in treated patients despite variable accepted that response evaluation criteria in solid tumors amounts of tumor shrinkage [1]. Antiangiogenic therapies are (RECIST) may be of limited utility in discriminating patients known to decrease tumor vascularization rather than result in responding to such therapies. direct cytotoxicity and have been associated with lesser Computed tomography (CT) is a robust, widely available degrees of tumor shrinkage than traditional antitumor agents, imaging modality routinely used in the assessment of patients’ with typical objective response rates ranging from 10% to 40% tumor burden throughout treatment and in monitoring Correspondence: Katherine M. Krajewski, M.D., Department of Imaging, Dana-Farber Cancer Institute, 450 Brookline Avenue, Boston, Massachusetts 02115, USA.Telephone: 617-582-8088; E-Mail: [email protected] Received October 10, 2013; accepted for publication February 21, 2014; first published online in The Oncologist Express on April 22, 2014. ©AlphaMed Press 1083-7159/2014/$20.00/0 http://dx.doi. org/10.1634/theoncologist.2013-0391 The Oncologist 2014;19:507–514 www.TheOncologist.com ©AlphaMed Press 2014 Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 508 10% Tumor Shrinkage and Outcome in Metastatic RCC patients for therapy-related toxicities. Several investigators patients had histologically confirmed stage IV mRCC. Patients have sought to optimize tumor response assessment via were treated at standard doses of study drug according to the imaging, such that imaging indicators better correlate with assigned protocol until they experienced disease progression, clinical benefit and survival outcomes early in the treatment severe or intolerable toxicity, or withdrew consent. Compliance course. Although the shortcomings of the conventional RECIST was checked after each cycle with a treatment diary. criteria in this settingareknown,there is no consensus asto the Imaging and Image Analysis preferred response assessment method. Several alternative Patients eligible for analysis included those with target lesions tumor shrinkage thresholds have been proposed, namely 10% by RECIST who underwent noncontrast-enhanced or contrast- and 20% long axis diameter reductions, either alone [7, 8] or in enhanced CTofthe chest, abdomen, and pelvisbefore andafter combination with other CT imaging response criteria, such as VEGF-targeted therapy initiation, with pre- and post-therapy attenuation [9–13]. Ten percent tumor shrinkage as an scans at the same institution.The routine oncology protocol was indicator of response has been supported in a study of patients used on multidetector CT scanners (64 detector row) using oral treated with sunitinib in a multicenter phase III trial [7] as contrast in all patients and intravenous contrast in patients with well as in another study of multiple imaging indicators of adequate estimated glomerular filtration rate (eGFR) and no response in patients treated with bevacizumab, sorafenib, known allergy to the contrast. Contrast-enhanced scans were and sunitinib [8]. Ten percent reduction in the long axis performed after 75–100 cc nonionic contrast administration diameter of target lesions is also a component of response (Ultravist; Bayer HealthCare, Leverkusen, Germany, http://www. accordingto Choi criteria [14], recently applied to patients with bayer.com;basedoneGFR),empiricallytimedwithchestimages mRCC [9, 12, 13]. Twenty percent long axis reductions are obtained in the arterial phase (30-second delay) and abdominal components of favorable response in the recently developed images obtained in the portal venous phase (70-second delay). size and attenuation CT (SACT) and morphology, attenuation, Images were reviewed and measured on a Picture Archiving and size, and structure (MASS) criteria [10, 11]. Communication System (Centricity; General Electric, Milwaukee, Inapreviousstudyof70mRCCpatientstreatedwithfirst-line WI, http://www.gehealthcare.com). sunitinib, sorafenib, or bevacizumab, four criteria of early Up to three target lesions in each patient were selected on postimaging changes were compared, including RECIST 1.0, 10% the baseline CT, according to RECIST 1.0 [20], by a single board- tumor diameter shrinkage, CT attenuation of tumor, and Choi certified radiologist with 8 years of experience in cancer criteria in assessing time-to-treatment failure (TTF) and overall imaging, blinded to patient outcomes. At baseline and first survival (OS) [8]. Ten percent tumor shrinkage at first follow-up follow-up, the longest axial axis of each target was recorded to was demonstrated statistically significant in predicting TTF and the nearest millimeter, as described previously [8]. Sum of the OS, whereas other criteria were not. However, given the longest diameter (SLD) of the targets was recorded at baseline retrospective nature of the study, it is necessary to validate the and at first follow-up of all patients. results in an independent cohort and reproduce the association RECIST response, including partial response (PR) and between 10% tumor shrinkage and survival, to propose use in progressive disease (PD), was recorded at CT follow-up, based future prospective clinical trials and clinical practice. on$30% decrease in SLD for PR and$20% increase in SLD for The purpose of the present study is to determine whether PD [20]. Patients not meeting either were tabulated as stable 10% tumor shrinkage can differentiate responders from disease (SD).Whereas RECISTresponse accordingto RECIST 1.0 nonresponders in an independent validation cohort of mRCC was studied in this report, the RECIST categorizations are patients enrolled in prospective clinical trials of vascular thought to be comparable to those according to the updated endothelial growth factor (VEGF) inhibitors. We hypothesize RECIST 1.1 criteria, because the definitions of PR, PD, and SD that10%tumorshrinkageatfirstfollow-up,asdemonstratedin are unchanged [21, 22]. Ten percent decrease in SLD was also our previous report and others [7, 8], represents a more examined as an independent predictor of outcome. clinically meaningful response assessment than RECIST 1.0. If Clinical outcome data were gathered by review of medical found reproducible, a 10% tumor shrinkage criterion could be records, including date of treatment failure (progression as widely applicable to mRCC patients, to identify patients likely determined by RECIST) and date of death. TTF was defined as to benefit from treatment early after initiation of therapy. time from treatment initiation until failure, including pro- gression and death. Patients who did not progress were PATIENTS AND METHODS censored on the date of last follow-up. OS was defined as the time from treatment initiation until death. Alive patients were Patients and Treatment censored on the date of last follow-up. The study sample consists of patients with metastatic renal cell carcinoma enrolled in six recent and ongoing multicenter, open- Statistical Analysis label studies of VEGF-targeted agents (vatalanib, sunitinib, Patients were classified into responders (PR) versus SD or PD, sorafenib,pazopanib,foretinib,andtivozanib)(3,15–19).Patients according to RECIST 1.0.The survival distribution for patients in were part of approved Institutional Review Board (IRB) pro- the two groups was estimated according to the Kaplan-Meier tocols for advanced renal cell carcinoma (RCC) at the institution method [23]. The difference in the survival distribution in which baseline and follow-up clinical data were prospectively between the groups was evaluated using the log-rank test. collected, and this study was approved by the IRB and is Health An analysis of RECIST 1.0 clinical benefit, PR or SD versus PD, Insurance Portability and Accountability Act compliant. Patients was not performed as only one patient had PD at the time of enrolled in these trials who had been included in a previous first follow-up, rendering this invalid. To explore any effect study cohort were excluded from this validation cohort [8]. All modifier among baseline characteristics in the relationships The ©AlphaMed Press 2014 Oncologist Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 Krajewski, Franchetti, Nishino et al. 509 Table 1. Patient demographic and disease characteristics of cohort had clear cell RCC histology (59%), 22 patients had papillary (33%), 2 patients each had chromophobe and the 66 patients unavailable histologic subtype, and 1 patient had predominant Characteristics Frequency Percent sarcomatoid features in addition to having clear cell histology. Age (median, range) 60.9 (32–84) The median age at targeted therapy initiation was 60.9 Gender years.Twenty-one patients (32%) were treated with vatalanib, 17 patients (26%) with foretinib, 10 patients (15%) each with Male 46 70 pazopanib and sunitinib, 6 patients (9%) with sorafenib, and 2 Female 20 30 patients (3%) with tivozanib. Median time from baseline CT to Histology drug initiation was 13 days, and median time from baseline to Clear cell 39 59 first post-treatment CT was 62 days (range 20–172 days). The Papillary 22 33 observed outlier of 172 days in time from baseline to first post- Chromophobe 2 3.0 treatment CT lay in 7 times the interquartile range (52 to 67 Sarcomatoid with clear cell 1 1.6 days) above the third quartile among all samples. All 66 Not specified 2 3.0 patients were evaluated according to RECIST 1.0 and 10% tumor shrinkage thresholds. Medication Using the selected target lesions for this study, most Vatalanib 21 31.8 patients (59 patients, 89%) were SD according to RECIST 1.0 at Foretinib 17 25.8 first post-treatment CT, whereas 6 were PR (9%) and 1 was PD Pazopanib 10 15.2 (2%). Using 10% tumor shrinkage to indicate response, 33 Sorafenib 6 9.1 patients achieved 10% tumor shrinkage and were considered Sunitinib 10 15.2 responders (50%; Fig. 1), whereas 33 patients achieved less Tivozanib 2 3.0 than 10% tumor shrinkage, deemed nonresponders. Pro- Reason for off study portionalchangesofSLD inreferencetobaselineinall but one of PD 38 57.6 the study patients at the time of first follow-up CT ranged from 52% decrease in SLD to 20% increase in the SLD of target lesions On trial 2 3.0 (Fig. 2). One patient with a single measurable target lesion in the Toxicity 24 36.4 peritoneum atbaselineexperienced significanttumorshrinkage Withdrew consent 2 3.0 such that the target lesion was no longer measurable at first CT size change from baseline 210.8 (2100, 20) follow-upCT,although nontargetascitesfluidremained present, to first follow-up CT (%) rendering the patient a PR according to RECIST. (median, range) No patient who achieved 10% tumor shrinkage at first Interval from baseline 13 [3, 27] CT to treatment initiation follow-up on retrospective assessment developed a new lesion (days) (median, range) or was deemed as having progressive disease at the time of the Interval from baseline CT to 62 (20, 172) original response assessment. post-treatment CT (days) (median, range) Association Between Early Post-Therapy Imaging Follow-up from registration 22.9 (1.3, 94.4) Changes and Clinical Outcome to the last documentation The TTF for patients with PR (median 6.9 months) was not or death (months) (median, range) significantly different from patients with SD or PD (median 5.5 Abbreviations: CT, computed tomography; PD, progressive disease. months, p5 .34; Fig. 3A). Similarly, the OS for patients with PR (median not reached, NR) was not significantly different from between decrease in SLD and survival outcomes, bivariate patients with SD or PD (median 25.5 months, p5 .72; Fig. 3B). analyses were performed using Cox proportional hazard Tumor shrinkage of $10% decrease in SLD significantly models in an exploratory fashion. differentiated responders (n 5 33, median TTF 8.4 months) Receiver-operating characteristics (ROC) analysis was used to from nonresponders (n 5 33, median TTF 4.1 months, p 5 determine the optimal threshold change in SLD at first CT follow- .001;Fig.4A).Tenpercentdecreasein SLDwasalsosignificantly uptodefinepatientswithsurvivalendpoint(deathorcensoredat predictive of overall survival (with responders OS of 35.1 the end of the study), similar to Thiam et al. [7]. After confirming months vs. nonresponders OS of 15.0 months, p 5 .003; validity of the area under the ROC curve (AUC) in an accuracy test Fig. 4B). Bivariate Cox proportional hazards models showed using a traditional point-system classification, Youden’sindex in that 10% decrease in SLD was a significantly robust predictor of the ROC analysis was calculated for the series of Ki-67 values TTF (p5 .001) and OS (p5 .01), respectively, after controlling observed in the data.We identified the threshold that minimized for either baseline age, gender, drug, histology, or Memorial the false-positive and false-negative rates for best detecting Sloan-Kettering Cancer Center risk (low/intermediate vs. prolonged survival, selecting patients with treatment benefit. high), whereas none of the controlled factors was significant. RESULTS Threshold Evaluation by ROC Analysis Patients For tumor shrinkage threshold evaluation, ROC curve analysis Sixty-six patients with 165 target lesions were eligible for the yielded29.3%inSLD(maximumYoudenindexscore0.28)asthe study and evaluated (Table 1). Thirty-nine patients of the optimal threshold for response/no response with respect to OS. www.TheOncologist.com ©AlphaMed Press 2014 Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 510 10% Tumor Shrinkage and Outcome in Metastatic RCC Figure 1. A 73-year-old man with metastatic renal cell carcinoma treated with foretinib. Axial contrast-enhanced computed tomography (CT) images at baseline (A and B) demonstrate target right adrenal and retroperitoneal metastases (black measurement lines), measuring 31 mm and 58 mm in long axis, respectively. Axial contrast-enhanced CT at first follow-up after treatment initiation (C and D) demonstrates approximately 10% decrease in the sum of the longest diameter of the targets (black measurement lines), measuring 28 mm and 52 mm, respectively. survival in 334 treated patients, including ROC analysis of DISCUSSION various other shrinkage thresholds [7]. The advocated 210% We evaluated RECISTresponse and tumor shrinkage as practical threshold in this study and the 230% threshold for PR and reproducible imaging predictors of benefit from various according to RECIST were reached after the first cycle of VEGF-targeted therapies in mRCC patients treated in recent and therapy in 73% and 19% of cases, respectively [7].Ten percent ongoingclinicaltrials.Inourstudyof66patients,asizablecohort tumor shrinkage was also a significant predictor of TTF and OS, with respect to similar literature, a 10% decrease in target SLD whereas RECIST and Choi criteria were not predictive in was a reliable predictor of TTF and OS outcomes on first follow- another study of patients treated with bevacizumab, sunitinib, up CT. This finding lends further support to the 10% tumor and sorafenib [8]. The current study lends further support to shrinkage threshold as an indicatorof response in this setting, as the applicability and utility of 10% tumor shrinkage as an early previously advocated in the literature [7, 8]. Such early changes indicatorof response in mRCC patients treatedwith avariety of in post-therapy imaging enable us to distinguish responder VEGF-targeted therapies, as class-specific tumor response patients and aid in patient management in three ways: (a) criteria rather than drug-specific. permitting patients who achieve this degree of tumor shrinkage Choi criteria have been applied to mRCC patients treated to continue treatment with greater confidence, (b) limiting with VEGF-targeted therapies, with differing conclusions on unnecessary toxicities to patients without evidence of treat- the applicability of these criteria in different reports. Choi ment benefit, and (c) enabling these patients to consider other criteria similarly incorporate 10% tumor shrinkage into “re- therapies. sponse,”or 15% decrease in mean tumor CT attenuation. In one The 10% tumor shrinkage threshold as an indicator of such report by van der Veldt et al. [9], Choi criteria better response to VEGF-targeted therapies in mRCC patients has delineated responders and nonresponders to therapy compared been supported by analysis ofa prior phase III studyofsunitinib with RECIST, but did not indicate early progression. Thereafter, in which 10% tumor shrinkage was the best predictor of Choicriteria were foundto earlydiscriminate patients benefiting The ©AlphaMed Press 2014 Oncologist Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 Krajewski, Franchetti, Nishino et al. 511 The unmet clinical need and utility of more generalized tumor response criteria are emphasized by various criteria evaluated in recent literature. A tumor response criterion that is generalizable as well as indicative of treatment benefit, such as one based on a tumor shrinkage threshold, is preferred and sought in this population. CT scanning is widely available and serves multiple purposes in the care of patients with widespread metastatic renal cell carcinoma. It is commonly used to assess the following: (a) patients’overall tumor burden intermsofgrowthpatterns anddistribution; (b)individual sites of metastatic disease to anticipate and limit troublesome disease-related complications via alternative, focused treat- ments (such as surgery or radiation to spinal metastases that may result in cord compression); and (c) evidence of treatment-related toxicities, potentially before the patient develops symptoms (such as pneumatosis intestinalis and/or hemorrhage in VEGF-targeted treatments) (25, 26). A simple, reliablemeasurementschemathathasabettercorrelationwith survival than RECIST can easily be incorporated into routine clinical practice at any center, regardless of scan protocol. Furthermore, single kidney patients with elevated eGFR or those with allergy may be assessed accordingly using the 10% Figure 2. Range of change in the sum long axis diameter of target tumorshrinkage, withoutthe necessityofintravenouscontrast. lesions at first follow-up computed tomography (CT). Responder patients with 10% tumor shrinkage at the time of the first follow- There are several limitations to this study. For instance, up CT (blue bars) had median time-to-treatment failure (TTF) and the imaging review was performed retrospectively, by a single overall survival (OS) of 8.4 and 35.1 months, respectively, whereas oncoradiologist, and patients were treated at a single in- nonresponder patients who did not achieve at least 10% tumor stitution. Given the small number of patients and the shrinkage (red bars) had median TTF and OS of 4.1 and 15.0 retrospective nature of the study, comparison of the 10% tumor months (p 5 .0014 and .0032, respectively). shrinkage threshold with the RECIST criteria is limited and should not be overinterpreted. In fact, in this study, we applied from treatment in at least one other report [12], although they the RECIST definitions of partial response, stable disease, and didnotsignificantlycorrelatewithsurvivaloutcomesinothers[8, progressive disease to the time of first follow-up CT rather than 13, 24]. In each of these studies, Choi criteria were more likely to assess RECIST response at all study time points. Of note, most characterizepatientsasrespondersthanothercriteriaingeneral, RECISTresponses do happen within 2 months (60 days median). regardless of outcome. Technical CT parameters, contrast, CT According to Thiam et al. [7], 64% of partial responses were attenuation measurementvariability, and patientvariables most achievedwithintwocyclesofsunitinibtreatment.Therefore,itis likely contribute to the differences in results observed. In likely that we captured most patients who achieved partial addition, no singular attenuation change threshold has been response forthe best response duringtheirtreatmentcourse,as found to correlate with survival [8], raising questions regarding a result of the timing of the first follow-up CTs. theapplicabilityofthe15%decreasedattenuationcomponentof Our purpose was to evaluate early changes in tumor burden this response schema in the mRCC population. on first CT follow-up and correlate these imaging changes with Similarly, the SACT and subsequent MASS criteria were survival. Although nadir tumor burden or “best overall re- developed to delineate response in mRCC patients treated with sponse” is important, this information is not available until sunitinib and sorafenib. These criteria include 20% tumor a patient has finished therapy. Early tumor shrinkage is useful shrinkage, marked central necrosis, or marked decreased because it can be assessed prospectively in all patients. Our attenuation as components of favorable response, whereas results show that early tumor shrinkage is informative, although 20% increase in size, new lesions, marked central fill-in, or new the results of our study should be interpreted in the context of enhancement are features of unfavorable response [10, 11]. the study design. Another persistent limitation in this validation PretherapyschemahavebeenrecentlycombinedwiththeMASS cohort is that it may be difficult to place patients close to the criteria to improve accuracy [24], offering a unique clinical and threshold (9% decrease) in the appropriate response category, imaging response model. However, drawbacks of these criteria but this would be true for any criteria using a threshold. includearbitrarycorrelationwithprogression-freesurvival.250 Measurement reproducibility in this cohort was separately days, exclusion of lung lesions, and volumetric measurements evaluated [27]. In this study, we found 10% tumor shrinkage was necessitating another workstation in SACT. In contrast to these a reproducible radiologic response indicator when baseline and methods, the 10% tumor shrinkage assessment method de- follow-up studies were measured by a single radiologist. scribed in this work may be advantageous because of the However, 10% tumor shrinkage was within the range of simplicity of single long axis diameter measurements of only interobserver measurement variability. The validation study three target lesions, which can be performed on a routineclinical findings must be considered in this context.Some of the benefits basiswithoutaddedtimeoraseparateadditionalworkstation,as of the patient cohort in this study include fairly homogeneous is needed in volumetric measurements. timing of the baseline and follow-up scans. Patients in this study www.TheOncologist.com ©AlphaMed Press 2014 Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 512 10% Tumor Shrinkage and Outcome in Metastatic RCC Figure 3. Time to treatment failure (TTF) and overall survival (OS) stratified by RECIST response at first follow-up computed tomography. (A): TTF stratified by RECIST response (PR) versus no response (SD 1 PD). TTF was not significantly different between RECIST responders and nonresponders (6.9 vs. 5.5 months, p 5 .3418). Partial response occurred in 6 patients at first follow-up, 6 events; 60 patients were nonresponders (SD 1 PD; 57 events and 3 censored). (B): OS stratified by RECIST response (PR) versus no response (SD 1 PD). OS was not significantly different between RECIST responders and nonresponders (not reached vs. 25.5 months, p5 .7185). Partial response occurred in 6 patients at first follow-up, with 3 patients deceased and 3 censored; 60 patients were nonresponders (SD1 PD; 43 deceased and 17 censored). Abbreviations: PD, progressive disease; PR, partial response; RECIST, Response Evaluation Criteria In Solid Tumors; SD, stable disease. The ©AlphaMed Press 2014 Oncologist Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 Krajewski, Franchetti, Nishino et al. 513 Figure 4. Time-to-treatment failure (TTF) and overall survival (OS) stratified by first follow-up computed tomography response according to .10%decreaseinsizeversusnoresponse.(A):TTFstratifiedbyresponseaccordingto$10%decreaseinsizeversusnoresponse(,10%decrease in size). TTF was significantly different between responders who achieved 10% decrease in size and nonresponders who did not (8.4 vs. 4.1 months, p 5 .0014). The $10% tumor shrinkage occurred in 33 patients at first follow-up, with 30 events and 3 censored; 33 patients were nonresponders(33events).(B):OSstratifiedbyresponseaccordingto$10%decreaseinsizeversusnoresponse(,10%decreaseinsize).OSwas significantlydifferent betweenresponders whoachieved10% decrease insizeand nonresponders whodidnot(35.1vs. 15.0months, p5.0032). The $10% tumor shrinkage occurred in 33 patients at first follow-up, with 18 patients deceased and 15 censored; 33 patients were nonresponders, with 28 patients deceased and 5 censored. Abbreviation: SLD, sum of the longest diameter. www.TheOncologist.com ©AlphaMed Press 2014 Downloaded from https://academic.oup.com/oncolo/article/19/5/507/6399359 by DeepDyve user on 01 February 2022 514 10% Tumor Shrinkage and Outcome in Metastatic RCC were on clinical trials of various VEGF-targeted agents, whereas AUTHOR CONTRIBUTIONS patientsinapreviousstudywerenottreatedontrials[8].Second, Conception/Design: Katherine M. Krajewski, Nikhil Ramaiya, Toni K. Choueiri data and conclusions from this study may be applicable to mRCC Provision of study material or patients: Toni K. Choueiri Collection and/or assembly of data: Katherine M. Krajewski, Mizuki Nishino, patients treated with VEGF-targeted therapies in general. Andre´ P. Fay, Toni K. Choueiri Data analysis and interpretation: Katherine M. Krajewski, Yoko Franchetti, CONCLUSION Mizuki Nishino, Andre´ P. Fay, Nikhil Ramaiya, Annick D.Van den Abbeele,Toni K. Choueiri The 10% tumor shrinkage threshold is predictive of survival in Manuscript writing: Katherine M. Krajewski, Yoko Franchetti, Mizuki Nishino, this validation cohort of mRCC patients treated with VEGF- Nikhil Ramaiya, Annick D. Van den Abbeele, Toni K. Choueiri Final approval of manuscript: Katherine M. Krajewski, Yoko Franchetti, Mizuki targeted therapies. ROC analysis identified 9.3% decrease in SLD Nishino, Andre P. Fay, Nikhil Ramaiya, Annick D. Van den Abbeele at first follow-up as the best size predictor of time-to-treatment failure.Thus, a reduction of 10% in the sum longaxis diameter of DISCLOSURES targets is validated as a reliable, early predictor of outcome in Katherine M. Krajewski: General Electric (RF); Toni K. Choueiri: Pfizer, this setting. Given the utility of this response indicator in Aveo, Exilixis, Novartis (C/A).The other authors indicated no financial multiple reports,the 10% tumorshrinkage may provide a simple relationships. (C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert and practical aid for therapeutic decision making if used in testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/ clinical trials and practice with a prospective evaluation. inventor/patent holder; (SAB) Scientific advisory board REFERENCES 1. Heng DY, Xie W, Regan MM et al. External Importance of size and attenuation on contrast- 19. Hutson TE, Rathmell K, Hudes GR et al. A phase validation and comparison with other models of the enhanced CT. AJR Am J Roentgenol 2010;194: II biomarker assessment of tivozanib in oncology International Metastatic Renal Cell Carcinoma 157–165. (BATON) trial in patients (pts) with advanced renal Database Consortium prognostic model: A popula- cell carcinoma (RCC). J Clin Oncol 2012;30(suppl): 11. Smith AD, Shah SN, Rini BI et al. Morphology, tion based study. Lancet Oncol 2013;14:141–148. abstract TPS4686. attenuation, size, and structure (MASS) criteria: 2. Yang JC, Haworth L, Sherry RM et al. A Assessing response and predicting clinical outcome 20. Therasse P, Arbuck SG, Eisenhauer EA et al. randomized trial of bevacizumab, an anti-vascular in metastatic renal cell carcinoma on antiangiogenic New guidelines to evaluate the response to endothelial growth factor antibody, for metastatic targeted therapy. AJR Am J Roentgenol 2010;194: treatment in solid tumors. European Organization 1470–1478. for Research and Treatment of Cancer, National renal cancer. N Engl J Med 2003;349:427–434. Cancer Institute of the United States, National 12. Schmidt N, Hess V, Zumbrunn T et al. Choi 3. Motzer RJ, Michaelson MD, Redman BG et al. Cancer Institute of Canada. J Natl Cancer Inst 2000; response criteria for prediction of survival in Activity of SU11248, a multitargeted inhibitor of 92:205–216. patients with metastatic renal cell carcinoma vascular endothelial growth factor receptor and treated with anti-angiogenic therapies. Eur Radiol 21. EisenhauerEA,TherasseP,BogaertsJ etal.New platelet-derived growth factor receptor, in patients 2013;23:632–639. response evaluation criteria in solid tumours: with metastatic renal cell carcinoma. J Clin Oncol Revised RECIST guideline (version 1.1). Eur J Cancer 2006;24:16–24. 13. Hittinger M, Staehler M, Schramm N et al. 2009;45:228–247. Course of size and density of metastatic renal cell 4. Motzer RJ, Hutson TE, Tomczak P et al. Sunitinib carcinoma lesions in the early follow-up of molec- 22. Nishino M, Jagannathan JP, Ramaiya NH et al. versus interferon alfa in metastatic renal-cell ular targeted therapy. Urol Oncol 2012;30:695–703. Revised RECIST guideline version 1.1: What oncol- carcinoma. N Engl J Med 2007;356:115–124. ogists want to know and what radiologists need to 14. Choi H, Charnsangavej C, Faria SC et al. 5. Escudier B, Eisen T, Stadler WM et al; TARGET know. AJR Am J Roentgenol 2010;195:281–289. Correlation of computed tomography and positron Study Group. Sorafenib in advanced clear-cell renal- emission tomography in patients with metastatic 23. Kaplan EL, Meier P. Nonparametric estimation cell carcinoma. N Engl J Med 2007;356:125–134. gastrointestinal stromal tumor treated at a single from incomplete observations. J Am Stat Assoc 6. Motzer RJ, Michaelson MD, Rosenberg J et al. institution with imatinib mesylate: Proposal of new 1958;53:457–481. Sunitinib efficacy against advanced renal cell computed tomography response criteria. J Clin 24. Smith AD, Shah SN, Rini BI et al. Utilizing pre- carcinoma. J Urol 2007;178:1883–1887. Oncol 2007;25:1753–1759. therapy clinical schema and initial CT changes to 7. Thiam R, Fournier LS, Trinquart L et al. Optimiz- 15. George D, Michaelson D, Oh WK et al. Phase I predict progression-free survival in patients with ing the size variation threshold for the CT evaluation study of PTK787/ZK 222584 (PTK/ZK) in metastatic metastatic renal cell carcinoma on VEGF-targeted of response in metastatic renal cell carcinoma renal cell carcinoma. Proc Am Soc Clin Oncol 2003; therapy: A preliminary analysis. Urol Oncol 2013;31: treated with sunitinib. Ann Oncol 2010;21:936–941. 22: abstract 1548. 1283–1291. 8. Krajewski KM, Guo M,Van den Abbeele AD et al. 16. Stadler WM, Figlin RA, McDermott DF et al; 25. Coriat R, Ropert S, Mir O et al. Pneumatosis Comparison of four early posttherapy imaging ARCCS Study Investigators. Safety and efficacy intestinalis associated with treatment of cancer changes (EPTIC; RECIST 1.0, tumor shrinkage, results of the advanced renal cell carcinoma patients with the vascular growth factor receptor computed tomography tumor density, Choi criteria) sorafenib expanded access program in North tyrosine kinase inhibitors sorafenib and sunitinib. in assessing outcome to vascular endothelial growth America. Cancer 2010;116:1272–1280. Invest New Drugs 2011;29:1090–1093. factor-targeted therapy in patients with advanced 17. MotzerRJ,Hutson TE,ReevesJetal.Randomized 26. Sonpavde G, Bellmunt J, Schutz F et al. The renal cell carcinoma. Eur Urol 2011;59:856–862. open-label phase III trial of pazopanib versus sunitinib double edged sword of bleeding and clotting from 9. van der Veldt AA, Meijerink MR, van den in first-line treatment of patients withmetastatic renal VEGF inhibition in renal cancer patients. Curr Oncol Eertwegh AJ et al. Choi response criteria for early cell carcinoma (MRCC): Results of the COMPARZ trial. Rep 2012;14:295–306. prediction of clinical outcome in patients with Paper presented at: 2012 ESMO Congress; October 1, 27. Krajewski KM, Nishino M, Franchetti Y et al. metastatic renal cell cancer treated with sunitinib. 2012, Abstract LBA8; Vienna, Austria. Intraobserver and interobserver variability in com- Br J Cancer 2010;102:803–809. 18. Choueiri TK, Vaishampayan U, Rosenberg JE puted tomography size and attenuation measure- 10. Smith AD, Lieber ML, Shah SN. Assessing tumor et al. Phase II and biomarker study of the dual MET/ ments in patients with renalcell carcinomareceiving response and detecting recurrence in metastatic VEGFR2 inhibitor foretinib in patients with papillary antiangiogenic therapy: Implications for alternative renal cell carcinoma on targeted therapy: renal cell carcinoma. J Clin Oncol 2013;31:181–186. response criteria. Cancer 2014;120:711–721. EDITOR’S NOTE: See the related commentary by Helen X. Chen et al. on page 439. The ©AlphaMed Press 2014 Oncologist

Journal

The Oncologist – Oxford University Press

Published: May 1, 2014

There are no references for this article.