Additional value of EUS in oesophageal cancer patients staged N0 on PET/CT: validation of a prognostic model

Additional value of EUS in oesophageal cancer patients staged N0 on PET/CT: validation of a... Background Lymph node metastases are a major prognostic indicator in oesophageal cancer. Radiological staging largely influences treatment decisions and is becoming more reliant on PET and CT. However, the sensitivity of these modalities is suboptimal and is known to under-stage disease. The primary aim of this study was to validate a published prognostic model in oesophageal cancer patients staged N0 with PET/CT, which showed that EUS nodal status was an independent predictor of survival. The secondary aim was to assess the prognostic significance of pathological lymph node metastases in this cohort. Methods An independent validation cohort included 139 consecutive patients from a regional upper gastrointestinal cancer network staged N0 with PET/CT between 1st January 2013 and 31st June 2015. Replicating the original study, two Cox regression models were produced: one included EUS T-stage and EUS N-stage, and one included EUS T-stage and EUS N0 versus N+. The primary outcome of the prognostic model was overall survival (OS). Kaplan–Meier analysis assessed differences in OS between pathological node-negative (pN0) and node-positive (pN+) groups. A p value of < 0.05 was considered statistically significant. Results The mean OS of the validation cohort was 29.8 months (95% CI 27.1–35.2). EUS T-stage was significantly and independently associated with OS in both models (p = 0.011 and p = 0.012, respectively). EUS N-stage and EUS N0 versus N+ were not significantly associated with OS (p = 0.553 and p = 0.359, respectively). There was a significant difference in OS between pN0 and pN+ groups (χ 13.315, df 1, p < 0.001). Conclusion Lymph node metastases have a significant detrimental effect on OS. This validation study did not replicate the results of the developed prognostic model but the continued benefit of EUS in patients staged N0 with PET/CT was demon- strated. EUS remains a valuable component of a multi-modality approach to oesophageal cancer staging. Keywords Oesophageal cancer · Endoscopic ultrasound · Positron-emission tomography · Prognosis · Survival In oesophageal cancer, clinical management is largely influ- therapy prior to surgery, routinely involving a combination enced by radiological lymph node staging. Accurate radio- of chemotherapy and radiotherapy [1]. logical staging is therefore vital to guide optimum treatment Under-staging of lymph node metastases increases the decisions for each patient. Patients considered to have lymph likelihood of recurrence and has a negative impact upon node metastases are more likely to receive neo-adjuvant overall survival (OS) [2]. Under-staging of disease can also impact on patient’s quality of life, potentially exposing them to high-risk surgical resection or definitive chemo-radiother - * K. G. Foley apy with toxic side-effects [3 ]. Currently, patients consid- foleykg@cardiff.ac.uk ered to be suitable for radical therapy undergo PET/CT and Division of Cancer & Genetics, School of Medicine, Cardiff EUS for further detailed assessment of local and distant dis- University, Cardiff CF14 4XN, UK ease [4]. However, EUS utilisation in the UK is declining. Wales Research and Diagnostic Positron Emission Data from the National Oesophago-Gastric Cancer Audit Tomography Imaging Centre (PETIC), Cardiff, UK (NOGCA) show 47.5% of patients had staging EUS in 2016 Department of Surgery, University Hospital of Wales, compared to 62% in 2013 [5]. Cardiff, UK As a result, there is an increasing reliance on cross-sec- Department of Radiology, University Hospital of Wales, tional imaging for treatment planning. The low sensitivity of Cardiff, UK Vol.:(0123456789) 1 3 4974 Surgical Endoscopy (2018) 32:4973–4979 PET/CT can increase the false-negative diagnostic rate for All patients in the validation cohort followed the usual staging lymph node metastases, under-staging regional disease and pathway and had PET/CT in site 2 using the same scanner and contributing to suboptimal treatment decisions [6]. protocol [7]. All patients had complete EUS staging using the A prognostic model investigating the additional role of same technique as described in the development cohort [7]. EUS in patients staged N0 on PET/CT was developed [7]. All staging was classified according to the TNM 7th edition The results of the study showed a significant difference in [9]. As in the previous study, 2 variables were recorded for OS in patients with EUS-positive nodal disease compared each patient: EUS T-stage (T1–4a) and EUS N-stage (N0–3). to EUS N0 disease. The inherently poor spatial resolution A third variable was derived from the EUS N-stage: EUS N0 of PET was thought to affect the differentiation of peri- versus N+ (N1, N2 or N3). tumoural nodes and the detection of small lymph node metastases, compared to the superior spatial resolution of Survival data EUS [8]. The primary aim of this study was to validate the prog- The primary outcome of the study was OS, defined in months nostic model in an independent cohort of oesophageal cancer from the date of diagnosis. Survival data were updated in July patients staged N0 with PET/CT. The secondary aim was 2016 ensuring at least 12 months of follow-up per patient. to assess the prognostic significance of pathological lymph Patients were followed-up every 3 months for the first year, node metastases in these patients. then every 6  months thereafter. No patients were lost to follow-up. Materials and methods Statistical analysis This was a validation study of a previously published prog- Continuous data were expressed as median (range) and cat- nostic model [7]. The setting was a regional upper gastroin- egorical data as frequency (%). Univariate analysis was per- testinal cancer network serving a population of 1.5 million. formed for EUS T-stage, EUS N-stage, and EUS N0 versus The prognostic model was developed in patients staged with N+, and differences between groups assessed using the log- PET/CT in two centres (sites 1 and 2). Validation was con- rank test [10]. Two Cox regression models were constructed to ducted in an independent cohort of patients staged N0 with assess the independent prognostic value of variables; model 1 PET/CT in site 2 only. Scientific review by the Research included EUS T-stage and EUS N-stage and model 2 included Review Board was performed and institutional research EUS T-stage and EUS N0 versus N+ [11]. Kaplan–Meier anal- approval was obtained (reference 13//DMD5769). Formal ysis using the log-rank test assessed differences in OS between ethical approval was not required for this study. the development and validation cohorts. In addition, model discrimination was assessed using a log-rank test to evaluate Development and validation cohorts OS differences between pN0 and pN+ groups in the sub-group of patients who underwent surgical resection in the valida- One-hundred and seventeen patients were included in the tion cohort. The effect of two different PET/CT scanners in development patient cohort. Consecutive patients were the development study was further investigated by excluding staged N0 with PET/CT between 1st December 2008 and patients from site 1 and re-calculating the models. The event- 31st May 2012. PET/CT examinations were performed per-variable (EPV) ratio ensured that the study was adequately across 2 sites; 47 in the first centre (site 1) and 70 in the powered, with a minimum EPV of 10 recommended, and an second centre (site 2). The PET/CT protocols were previ- event defined as a death [12]. A p value of < 0.05 was con- ously published [7]. Patients with M1 disease (non-regional sidered statistically significant. Statistical analysis was per - nodal disease or distant metastases) on PET/CT (n = 6) or formed with SPSS v23 (IBM, Chicago, USA). This validation those with incomplete EUS staging (n = 39) were excluded. study is reported according to the Transparent Reporting of a All patients were staged N0 on PET/CT by a Consultant Multivariable Prediction Model for Individual Prognosis or Radiologist certified in PET/CT reporting and had staging Diagnosis (TRIPOD) guidelines [13]. EUS completed. All staging was classified according to the TNM 7th edition [9]. The same selection criteria were applied to the validation Results cohort. Patients staged N0 with PET/CT between 1st January 2013 and 31st June 2015 were considered for the validation Patient cohorts cohort (n = 166). Patients with distant metastases (n = 4) or incomplete EUS staging (n = 23) were excluded. Following The baseline characteristics of development and validation exclusions, 139 patients were included in the validation study. patient cohorts are included in Table  1. There were no 1 3 Surgical Endoscopy (2018) 32:4973–4979 4975 missing data in this study. The median age of the devel- Summary of results from development cohort opment and validation cohorts was 67.0  years (range 24.0–82.0) and 66.0  years (39–84), respectively. The Univariate analysis showed that EUS T-stage (Χ 8.321, median follow-up period was 25.0 months in the develop- df 3, p = 0.040), N-st age (Χ 14.879, df 3, p = 0.002), ment cohort (95% CI 23.1–26.9) and 26.0 months (95% and N0 versus N+ (χ 11.325, df 1, p = 0.001) were sig- CI 22.7–29.3) in the validation cohort. Mean survival nificantly associated with OS. When EUS T-stage and times are presented because a 50% mortality rate was not N-stage were entered in model 1, only EUS N-stage was reached in either cohort. The mean OS of the develop- significantly and independently associated with dura- ment cohort was 33.1  months (95% CI 30.1–36.1) and tion of survival [hazard ratio (HR) 1.616–4.707, 95% CI 29.8 months (95% CI 27.1–35.2) in the validation cohort. 0.363–7.190, p = 0.005]. When EUS T-stage and EUS N0 versus N+ were entered in model 2, EUS N0 versus N+ Table 1 Baseline characteristics Frequency (%) Development cohort (n = 117) Validation cohort (n = 139) of patients in development and validation cohorts Male:female 88 (75.2):29 (24.8) 108 (77.7):31 (22.3) Tumour location  Oesophagus 73 (62.4) 76 (54.7)   Upper 0 (0.0) 2 (2.7)   Mid 20 (27.4) 22 (28.9)   Distal 53 (72.6) 52 (68.4)  Junction 44 (37.6) 63 (45.3)   Siewert type I 5 (11.3) 25 (39.7)   Siewert type II 12 (27.3) 18 (28.6)   Siewert type III 27 (61.4) 20 (31.7)  Histology   Adenocarcinoma 98 (83.8) 107 (77.0)   SCC 19 (16.2) 26 (18.7)   HGD 0 (0.0) 3 (2.2)   Neuro-endocrine 0 (0.0) 2 (1.4)   Undifferentiated 0 (0.0) 1 (0.7)  EUS T-stage   T1 18 (15.4) 20 (14.4)   T2 16 (13.7) 18 (12.9)   T3 75 (64.1) 86 (61.9)   T4a 8 (6.8) 13 (9.4)   T4b 0 (0.0) 2 (1.4) EUS N-stage   N0 78 (66.7) 89 (64.0)   N1 23 (19.7) 42 (30.2)   N2 9 (7.6) 7 (5.1)   N3 7 (6.0) 1 (0.7) Treatment  Curative 105 (89.7) 116 (83.5)   NACT 40 (38.1) 44 (37.9)   dCRT 29 (27.6) 39 (33.6)   Surgery alone 32 (30.5) 19 (16.4)   NACRT 1 (0.9) 11 (9.5)   EMR 3 (2.9) 3 (2.6)  Palliative 12 (10.3) 23 (16.5)  Mortality   Alive 84 (71.8) 85 (61.2)   Dead 33 (28.2) 54 (38.8) 1 3 4976 Surgical Endoscopy (2018) 32:4973–4979 Table 2 Survival data of patients in validation cohort derived from univariate analysis EUS variable Mean OS (months) 95% confidence interval Lower Upper T-stage  T1 41.563 38.834 44.291  T2 25.830 20.062 31.598  T3 27.908 24.484 31.332  T4a 16.846 9.842 23.851  T4b 17.500 4.334 30.666 N-stage  N0 31.853 28.735 34.971  N1 25.625 21.246 30.004  N2 16.857 11.873 21.841  N3 17.000 17.000 17.000 Fig. 1 Significant difference in cumulative survival by EUS T-stage  N+ 24.924 20.966 28.882 in Validation cohort (χ 21.031, df 4, p < 0.001). Patients with more advanced EUS T-stage have worse OS Table 3 Results of model 1 multi-variate analysis including EUS was significantly and independently associated with OS T-stage and EUS N-stage in validation cohort (HR 3.105, 95% CI 1.543–6.247, p = 0.001). Variable p value Hazard ratio df 95% confidence interval Univariate analysis in validation cohort Lower Upper EUS T-stage 0.011 4 EUS T-stage (χ 21.031, df 4, p < 0.001) (Fig. 1) and EUS  T2 0.018 12.482 1 1.528 101.957 N0 versus N+ (χ 4.300, df 1, p = 0.038) were significantly  T3 0.016 11.656 1 1.570 86.548 associated with OS. EUS N-stage did not show a statis-  T4a 0.001 30.114 1 3.731 243.079 tically significant association with OS (χ 5.699, df 3,  T4b 0.050 16.270 1 0.994 266.273 p = 0.127). Table 2 shows mean survival data for patients EUS N-stage 0.553 3 classified by EUS T-stage, N-stage, and EUS N0 versus  N1 0.560 1.192 1 0.660 2.154 N+.  N2 0.353 1.653 1 0.572 4.772  N3 0.260 3.176 1 0.425 23.716 Multivariate analysis in validation cohort Table 4 Results of model 2 multi-variate analysis including EUS Again, two Cox regression models were produced in the T-stage and EUS N0 versus N+ in validation cohort validation cohort. EUS T-stage and EUS N-stage were entered in model 1. EUS T-stage was significantly and inde - Variable p value Hazard ratio df 95% confidence interval pendently associated with OS (HR 11.656–30.114, 95% CI 0.994–243.079, p = 0.011). (Table  3) The EPV ratio was Lower Upper 27.0. EUS T-stage 0.012 4 EUS T-stage and EUS N0 versus N+ were entered in  T2 0.020 11.977 1 1.469 97.620 model 2. Again, EUS T-stage was significantly and inde-  T3 0.016 11.714 1 1.579 86.902 pendently associated with OS (HR 11.714–29.631, 95% CI  T4a 0.001 29.631 1 3.674 238.959 1.067–238.959, p = 0.012). (Table  4) The EPV ratio was  T4b 0.045 17.243 1 1.067 278.714 27.0. EUS N0 vs N+ 0.359 1.292 1 0.747 2.235 There was no statistically significant difference in OS between the development (mean OS 33.0 months, 95% CI 30.060–36.076) and validation cohorts (mean OS 29.8 months, 95% CI 27.120–32.513) (χ 1.979, df 1, p = 0.159). 1 3 Surgical Endoscopy (2018) 32:4973–4979 4977 Eec ff t of including site 1 patients in development cohort A post hoc analysis was performed to determine the effect of including patients scanned in site 1 on the development cohort prognostic models. To perform this post hoc analysis, site 1 patients were excluded from the development cohort in an attempt to control comparison with the validation cohort. Seventy patients were originally scanned at site 2. Of these, 53 patients (75.7%) were staged EUS N0; 11 (15.7%) were EUS N1; 3 (4.3%) were N2, and 3 (4.3%) were N3. Both EUS N-stage (HR 2.365–32.757, 95% CI 0.476–223.922, p = 0.005) and EUS N0 versus N+ (HR 3.783, 1.141–12.539, p = 0.030) were independent predictors of OS, in keeping with findings from the original study. Therefore, inclusion of site 1 patients had little effect on the prognostic models of the development cohort. Confidence intervals are wide, Fig. 2 Significant difference in cumulative survival depending on the likely due to the small numbers in N2 and N3 groups. presence of pathological lymph nodes (χ 13.315, df 1, p < 0.001). Patients with positive pathological lymph nodes have worse OS Comparison of patient eligibility for development and validation cohorts (p = 0.038). Relatively small numbers of patients staged EUS N2 and N3 were included in the development cohort A comparison of the proportion of patients who were staged (n = 9 and n = 7, respectively) which could result in dispro- N0 and considered for inclusion during both study periods portionate statistical significance and failure to validate this was made. Post hoc review revealed that 117 of 207 patients variable. Importantly, this study showed that EUS T-stage is (56.5%) from site 2 were staged N0 on PET/CT during the significantly and independently associated with OS, which study recruitment period of the development cohort and 139 supports other studies [14–16]. The study adds evidence to of 317 (43.8%) from site 2 were staged N0 during the study the importance of EUS in the multi-modality approach to recruitment period of the validation cohort. This difference oesophageal cancer staging. was statistically significant (χ 8.049, df 1, p = 0.005). A When patients scanned in site 1 were removed from the significantly higher proportion of patients from site 2 were development cohort, EUS N-stage and EUS N0 versus N+ staged N0 in the development cohort, suggesting that dif- remained independent predictors of OS. This finding sug- ferent proportions of patients were eligible for inclusion, gests that the site 1 PET/CT scanner had little effect on the potentially affecting the results of the prognostic models. developed prognostic model. One reason for the inability to validate the model could be the statistically significant Prognostic significance of pathological lymph nodes difference in proportions of patients staged N0 during both study periods, resulting in fewer patients from site 2 eligi- In total, 74 patients from the validation cohort underwent ble for inclusion in the validation study recruitment period surgical resection. Thirty-nine patients (52.7%) were classi- (those staged N0 on PET/CT). fied pN0 and 35 patients (47.3%) were classified pN+. There An important issue in validation studies is the extent to was a significant difference in OS between pN0 and pN+ which the included cohorts differ. This can result in valida- groups (χ 13.315, df 1, p < 0.001). (Fig. 2) Mean OS for the tion failure unless appropriately adjusted for [17]. Report- pN0 group was 40.091 months (95% CI 36.931–43.251) and ing trends over time have not been assessed in this study, 26.538 (95% CI 22.123–30.953) for the pN+ group. but reporting and context bias have been shown to influence radiologists image interpretation [18]. It is possible that a priori knowledge of key findings from the developed prog- Discussion nostic model may increase the likelihood of equivocal lymph nodes on PET/CT being called positive. This awareness of This validation study failed to replicate the results of the having been studied with consequent impact on behaviour, developed prognostic model. In the validation cohort, EUS the so-called ‘Hawthorne effect’, could have contributed to N-stage and EUS N0 versus N+ did not show prognostic the change in results [19]. However, this is merely a hypoth- significance in multi-variate analysis, although EUS N0 ver - esis and cannot be concluded from these data. In addition, sus N+ was statistically significant in univariate analysis 1 3 4978 Surgical Endoscopy (2018) 32:4973–4979 operators were not fully blinded to the results of the PET/CT, importance of rigorous study design and statistical analy- potentially influencing the interpretation of EUS N-stage. sis cannot be understated, and further research is required The final report of the PET/CT is routinely checked prior to improve prognostic research methodology [24, 25]. In to EUS to ensure that no distant metastases are detected, addition, it is important to document and publish the find - preventing inappropriate EUS examination. ings of all prognostic research, including ‘non-significant’ Prior to the opening of site 2 in 2010, patients were findings, since publication bias is prevalent [26]. scanned at site 1 using a Philips 16 section Gemini GXL This validation study is limited by some factors. Both dedicated PET/CT system (Philips Medical Systems, Cleve- development and validation cohorts represent a relatively land, USA). The two sites used different scanners and proto- heterogeneous cohort of patients; however, this is reflec- cols, and patients were scanned at different activity uptake tive of the demographics of oesophageal cancer patients times. Patients were scanned at 60 min of uptake time in in general. Dissemination of lymph node metastases is site 1 and after 90 min in site 2. Longer uptake times lead dependent on T-stage and to a lesser extent, histological to higher tumour to background avidity and can therefore cell type of the primary tumour [27, 28]. Stratification of increase the conspicuity of lymph node metastases. Sec- these factors was not performed, but most patients pre- ondly, the scanner in site 2 had a time-of-flight (TOF) algo- sented with locally advanced T3 or T4 tumours. The EUS rithm but the site 1 scanner did not. TOF reconstructions operators may have been influenced by results of the PET/ improve signal-to-noise ratio, detection and anatomical CT reports, resulting in inadvertent changes in lymph node localisation of lymph node metastases by allowing more assessment over time. In addition, the inclusion of patients precise measurement of the time difference between detec- from site 1 could not be replicated in the validation cohort tions [20]. Finally, images were acquired for 4 min per bed because our patients no longer attend this site for PET/CT. position in site 1, whereas the acquisition was 3 min per bed Despite the limitations, this validation study has sev- position in site 2. Some improvement in image quality may eral strengths. All patients were managed by an experi- be expected in site 1 with longer acquisition times, provided enced Upper gastrointestinal cancer multi-disciplinary the patient remained still. However, the results of this valida- team covering a population of approximately 1.5 million. tion study assume that longer acquisition did not affect the The site 2 PET/CT scanner and acquisition protocol was models. Standardised PET acquisition protocols have been unchanged in both development and validation cohorts. discussed in the literature [21], but these findings suggest the The study was adequately powered according to a recom- differences may be less influential than previously thought, mended EPV ratio. In addition, a pathologist reviewed the adding generalisability to the results. resected lymph nodes to evaluate and confirm the presence An additional hypothesis for the failure of validation is of pathological lymph node metastases. the accuracy of the staging investigations. Results from our institution have shown suboptimal diagnostic accuracy in the general oesophageal cancer population [22]. Overall accuracy of differentiating negative and positive nodal dis- Conclusion ease on CT, EUS, and PET/CT was 54.5, 55.4, and 57.1%, respectively. Sensitivity and specificity were 39.7 and 77.3% Validation studies are important in prognostic research with CT, 42.6 and 75.0% with EUS and 35.3 and 90.9% with [29]. Lymph node metastases have a significant detrimen- PET/CT. These results were largely attributable to the detec- tal effect on OS. This validation study did not replicate the tion of micro-metastases in 44% of lymph nodes (defined results of the developed prognostic model but the contin- as ≤ 2 mm) evaluated pathologically. The suboptimal accu- ued benefit of EUS in patients staged N0 with PET/CT racy could affect results of the developed prognostic model, was demonstrated. EUS remains a valuable component of obtaining a significant difference in survival between EUS a multi-modality approach to oesophageal cancer staging. N-stage categories by chance alone. Funding No funding was received for this study. The additional sub-group analysis conducted in this study confirms the presence of lymph node metastases as Compliance with ethical standards a major prognostic indicator [2]. There was a highly sig- nificant difference in OS between pN0 and pN+ groups in Disclosures Dr. Kieran Foley, Dr. Patrick Fielding, Professor Wyn patients staged N0 on PET/CT. This finding highlights the Lewis, and Dr. Ashley Roberts have no conflicts of interest and or importance of accurate pre-treatment lymph node staging in financial ties to disclose. oesophageal cancer. Model validation is not commonly performed in prog- Open Access This article is distributed under the terms of the Crea- nostic research. Approximately 34% of models are vali- tive Commons Attribution 4.0 International License (http://creat dated, with 11% undergoing external validation [23]. The iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, 1 3 Surgical Endoscopy (2018) 32:4973–4979 4979 distribution, and reproduction in any medium, provided you give appro- margin involvement following oesophagectomy for cancer and priate credit to the original author(s) and the source, provide a link to the predictive role of endoluminal ultrasonography. Br J Cancer the Creative Commons license, and indicate if changes were made. 107:1925–1931 16. Hiele M, DeLeyn P, Schurmans P, Lerut A, Huys S, Geboes K, Gevers AM, Rutgeerts P (1997) Relation between endoscopic ultrasound findings and outcome of patients with tumors of the References esophagus or esophagogastric junction. Gastrointest Endosc 45:381–386 1. Mukherjee S, Hurt CN, Gwynne S, Bateman A, Gollins S, Rad- 17. Debray TP, Vergouwe Y, Koffijberg H, Nieboer D, Steyerberg EW, hakrishna G, Hawkins M, Canham J, Lewis W, Grabsch HI, Moons KG (2015) A new framework to enhance the interpretation Sharma RA, Wade W, Maggs R, Tranter B, Roberts A, Sebag- of external validation studies of clinical prediction models. J Clin Montefiore D, Maughan T, Griffiths G, Crosby T (2015) NEO- Epidemiol 68:279–289 SCOPE: a randomised phase II study of induction chemotherapy 18. Egglin TK, Feinstein AR (1996) Context bias. A problem in diag- followed by either oxaliplatin/capecitabine or paclitaxel/carbopl- nostic radiology. JAMA 276:1752–1755 atin based chemoradiation as pre-operative regimen for resectable 19. McCambridge J, Witton J, Elbourne DR (2014) Systematic oesophageal adenocarcinoma. BMC Cancer 15:48 review of the Hawthorne effect: new concepts are needed to study 2. Kayani B, Zacharakis E, Ahmed K, Hanna GB (2011) Lymph research participation effects. J Clin Epidemiol 67:267–277 node metastases and prognosis in oesophageal carcinoma—a sys- 20. Surti S (2015) Update on time-of-flight PET imaging. J Nucl Med tematic review. Eur J Surg Oncol 37:747–753 56:98–105 3. Blazeby JM, Farndon JR, Donovan J, Alderson D (2000) A 21. Boellaard R, O’Doherty MJ, Weber WA, Mottaghy FM, Lonsdale prospective longitudinal study examining the quality of life of MN, Stroobants SG, Oyen WJ, Kotzerke J, Hoekstra OS, Pruim J, patients with esophageal carcinoma. Cancer 88:1781–1787 Marsden PK, Tatsch K, Hoekstra CJ, Visser EP, Arends B, Ver- 4. Roberts SA, Kay C (2014) Oesophagus and stomach cancers. In: zijlbergen FJ, Zijlstra JM, Comans EF, Lammertsma AA, Paans Nicholson T (ed) Recommendations for cross-sectional imaging AM, Willemsen AT, Beyer T, Bockisch A, Schaefer-Prokop C, in cancer management. Royal College of Radiologists, London Delbeke D, Baum RP, Chiti A, Krause BJ (2010) FDG PET and 5. National Oesophago-Gastric Cancer Audit (2016) An audit of the PET/CT: EANM procedure guidelines for tumour PET imaging: care received by people with oesophago-gastric cancer in Eng- version 1.0. Eur J Nucl Med Mol Imaging 37:181–200 land and Wales annual report. National Oesophago-Gastric Cancer 22. Foley KG, Christian A, Fielding P, Lewis WG, Roberts SA (2017) Audit, London Accuracy of contemporary oesophageal cancer lymph node 6. van Vliet EP, Heijenbrok-Kal MH, Hunink MG, Kuipers EJ, staging with radiological-pathological correlation. Clin Radiol Siersema PD (2008) Staging investigations for oesophageal can- 72:e691–e697 cer: a meta-analysis. Br J Cancer 98:547–557 23. Mallett S, Royston P, Waters R, Dutton S, Altman DG (2010) 7. Foley KG, Lewis WG, Fielding P, Karran A, Chan D, Blake P, Reporting performance of prognostic models in cancer: a review. Roberts SA (2014) N-staging of oesophageal and junctional car- BMC Med 8:21 cinoma: is there still a role for EUS in patients staged N0 at PET/ 24. Riley RD, Hayden JA, Steyerberg EW, Moons KG, Abrams K, CT? Clin Radiol 69:959–964 Kyzas PA, Malats N, Briggs A, Schroter S, Altman DG, Hem- 8. Kapoor V, McCook BM, Torok FS (2004) An introduction to ingway H, Group P (2013) Prognosis Research Strategy (PRO- PET-CT imaging. Radiographics 24:523–543 GRESS) 2: prognostic factor research. PLoS Med 10:e1001380 9. Sobin LH, Gospodarowicz MK, Wittekind CH (2009) UICC TNM 25. Moons KG, Royston P, Vergouwe Y, Grobbee DE, Altman DG classification of malignant tumours, 7th edn. Wiley, New York (2009) Prognosis and prognostic research: what, why, and how? 10. Kaplan E, Meier P (1958) Nonparametric estimation from incom- BMJ 338:b375 plete observations. J Am Stat Assoc 53:457–481 26. Kyzas PA, Denaxa-Kyza D, Ioannidis JP (2007) Almost all arti- 11. Cox DR (1972) Regression models and life-tables. J R Stat Soc cles on cancer prognostic markers report statistically significant 34:187–220 results. Eur J Cancer 43:2559–2579 12. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR (1996) 27. Siewert JR, Stein HJ, Feith M, Bruecher BL, Bartels H, Fink U A simulation study of the number of events per variable in logistic (2001) Histologic tumor type is an independent prognostic param- regression analysis. J Clin Epidemiol 49:1373–1379 eter in esophageal cancer: lessons from more than 1,000 consecu- 13. Collins GS, Reitsma JB, Altman DG, Moons KG (2015) Transpar- tive resections at a single center in the Western world. Ann Surg ent reporting of a multivariable prediction model for individual 234:360–369 prognosis or diagnosis (TRIPOD): the TRIPOD statement. BMJ 28. Rice TW, Zuccaro G Jr, Adelstein DJ, Rybicki LA, Blackstone 350:g7594 EH, Goldblum JR (1998) Esophageal carcinoma: depth of tumor 14. Paraf F, Flejou JF, Pignon JP, Fekete F, Potet F (1995) Surgi- invasion is predictive of regional lymph node status. Ann Thorac cal pathology of adenocarcinoma arising in Barrett’s esophagus. Surg 65:787–792 Analysis of 67 cases. Am J Surg Pathol 19:183–191 29. Altman DG, Royston P (2000) What do we mean by validating a 15. Reid TD, Chan DS, Roberts SA, Crosby TD, Williams GT, Lewis prognostic model? Stat Med 19:453–473 WG (2012) Prognostic significance of circumferential resection 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Surgical Endoscopy Springer Journals

Additional value of EUS in oesophageal cancer patients staged N0 on PET/CT: validation of a prognostic model

Free
7 pages

Loading next page...
 
/lp/springer_journal/additional-value-of-eus-in-oesophageal-cancer-patients-staged-n0-on-KUU0EnS038
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s)
Subject
Medicine & Public Health; Surgery; Gynecology; Gastroenterology; Hepatology; Proctology; Abdominal Surgery
ISSN
0930-2794
eISSN
1432-2218
D.O.I.
10.1007/s00464-018-6259-0
Publisher site
See Article on Publisher Site

Abstract

Background Lymph node metastases are a major prognostic indicator in oesophageal cancer. Radiological staging largely influences treatment decisions and is becoming more reliant on PET and CT. However, the sensitivity of these modalities is suboptimal and is known to under-stage disease. The primary aim of this study was to validate a published prognostic model in oesophageal cancer patients staged N0 with PET/CT, which showed that EUS nodal status was an independent predictor of survival. The secondary aim was to assess the prognostic significance of pathological lymph node metastases in this cohort. Methods An independent validation cohort included 139 consecutive patients from a regional upper gastrointestinal cancer network staged N0 with PET/CT between 1st January 2013 and 31st June 2015. Replicating the original study, two Cox regression models were produced: one included EUS T-stage and EUS N-stage, and one included EUS T-stage and EUS N0 versus N+. The primary outcome of the prognostic model was overall survival (OS). Kaplan–Meier analysis assessed differences in OS between pathological node-negative (pN0) and node-positive (pN+) groups. A p value of < 0.05 was considered statistically significant. Results The mean OS of the validation cohort was 29.8 months (95% CI 27.1–35.2). EUS T-stage was significantly and independently associated with OS in both models (p = 0.011 and p = 0.012, respectively). EUS N-stage and EUS N0 versus N+ were not significantly associated with OS (p = 0.553 and p = 0.359, respectively). There was a significant difference in OS between pN0 and pN+ groups (χ 13.315, df 1, p < 0.001). Conclusion Lymph node metastases have a significant detrimental effect on OS. This validation study did not replicate the results of the developed prognostic model but the continued benefit of EUS in patients staged N0 with PET/CT was demon- strated. EUS remains a valuable component of a multi-modality approach to oesophageal cancer staging. Keywords Oesophageal cancer · Endoscopic ultrasound · Positron-emission tomography · Prognosis · Survival In oesophageal cancer, clinical management is largely influ- therapy prior to surgery, routinely involving a combination enced by radiological lymph node staging. Accurate radio- of chemotherapy and radiotherapy [1]. logical staging is therefore vital to guide optimum treatment Under-staging of lymph node metastases increases the decisions for each patient. Patients considered to have lymph likelihood of recurrence and has a negative impact upon node metastases are more likely to receive neo-adjuvant overall survival (OS) [2]. Under-staging of disease can also impact on patient’s quality of life, potentially exposing them to high-risk surgical resection or definitive chemo-radiother - * K. G. Foley apy with toxic side-effects [3 ]. Currently, patients consid- foleykg@cardiff.ac.uk ered to be suitable for radical therapy undergo PET/CT and Division of Cancer & Genetics, School of Medicine, Cardiff EUS for further detailed assessment of local and distant dis- University, Cardiff CF14 4XN, UK ease [4]. However, EUS utilisation in the UK is declining. Wales Research and Diagnostic Positron Emission Data from the National Oesophago-Gastric Cancer Audit Tomography Imaging Centre (PETIC), Cardiff, UK (NOGCA) show 47.5% of patients had staging EUS in 2016 Department of Surgery, University Hospital of Wales, compared to 62% in 2013 [5]. Cardiff, UK As a result, there is an increasing reliance on cross-sec- Department of Radiology, University Hospital of Wales, tional imaging for treatment planning. The low sensitivity of Cardiff, UK Vol.:(0123456789) 1 3 4974 Surgical Endoscopy (2018) 32:4973–4979 PET/CT can increase the false-negative diagnostic rate for All patients in the validation cohort followed the usual staging lymph node metastases, under-staging regional disease and pathway and had PET/CT in site 2 using the same scanner and contributing to suboptimal treatment decisions [6]. protocol [7]. All patients had complete EUS staging using the A prognostic model investigating the additional role of same technique as described in the development cohort [7]. EUS in patients staged N0 on PET/CT was developed [7]. All staging was classified according to the TNM 7th edition The results of the study showed a significant difference in [9]. As in the previous study, 2 variables were recorded for OS in patients with EUS-positive nodal disease compared each patient: EUS T-stage (T1–4a) and EUS N-stage (N0–3). to EUS N0 disease. The inherently poor spatial resolution A third variable was derived from the EUS N-stage: EUS N0 of PET was thought to affect the differentiation of peri- versus N+ (N1, N2 or N3). tumoural nodes and the detection of small lymph node metastases, compared to the superior spatial resolution of Survival data EUS [8]. The primary aim of this study was to validate the prog- The primary outcome of the study was OS, defined in months nostic model in an independent cohort of oesophageal cancer from the date of diagnosis. Survival data were updated in July patients staged N0 with PET/CT. The secondary aim was 2016 ensuring at least 12 months of follow-up per patient. to assess the prognostic significance of pathological lymph Patients were followed-up every 3 months for the first year, node metastases in these patients. then every 6  months thereafter. No patients were lost to follow-up. Materials and methods Statistical analysis This was a validation study of a previously published prog- Continuous data were expressed as median (range) and cat- nostic model [7]. The setting was a regional upper gastroin- egorical data as frequency (%). Univariate analysis was per- testinal cancer network serving a population of 1.5 million. formed for EUS T-stage, EUS N-stage, and EUS N0 versus The prognostic model was developed in patients staged with N+, and differences between groups assessed using the log- PET/CT in two centres (sites 1 and 2). Validation was con- rank test [10]. Two Cox regression models were constructed to ducted in an independent cohort of patients staged N0 with assess the independent prognostic value of variables; model 1 PET/CT in site 2 only. Scientific review by the Research included EUS T-stage and EUS N-stage and model 2 included Review Board was performed and institutional research EUS T-stage and EUS N0 versus N+ [11]. Kaplan–Meier anal- approval was obtained (reference 13//DMD5769). Formal ysis using the log-rank test assessed differences in OS between ethical approval was not required for this study. the development and validation cohorts. In addition, model discrimination was assessed using a log-rank test to evaluate Development and validation cohorts OS differences between pN0 and pN+ groups in the sub-group of patients who underwent surgical resection in the valida- One-hundred and seventeen patients were included in the tion cohort. The effect of two different PET/CT scanners in development patient cohort. Consecutive patients were the development study was further investigated by excluding staged N0 with PET/CT between 1st December 2008 and patients from site 1 and re-calculating the models. The event- 31st May 2012. PET/CT examinations were performed per-variable (EPV) ratio ensured that the study was adequately across 2 sites; 47 in the first centre (site 1) and 70 in the powered, with a minimum EPV of 10 recommended, and an second centre (site 2). The PET/CT protocols were previ- event defined as a death [12]. A p value of < 0.05 was con- ously published [7]. Patients with M1 disease (non-regional sidered statistically significant. Statistical analysis was per - nodal disease or distant metastases) on PET/CT (n = 6) or formed with SPSS v23 (IBM, Chicago, USA). This validation those with incomplete EUS staging (n = 39) were excluded. study is reported according to the Transparent Reporting of a All patients were staged N0 on PET/CT by a Consultant Multivariable Prediction Model for Individual Prognosis or Radiologist certified in PET/CT reporting and had staging Diagnosis (TRIPOD) guidelines [13]. EUS completed. All staging was classified according to the TNM 7th edition [9]. The same selection criteria were applied to the validation Results cohort. Patients staged N0 with PET/CT between 1st January 2013 and 31st June 2015 were considered for the validation Patient cohorts cohort (n = 166). Patients with distant metastases (n = 4) or incomplete EUS staging (n = 23) were excluded. Following The baseline characteristics of development and validation exclusions, 139 patients were included in the validation study. patient cohorts are included in Table  1. There were no 1 3 Surgical Endoscopy (2018) 32:4973–4979 4975 missing data in this study. The median age of the devel- Summary of results from development cohort opment and validation cohorts was 67.0  years (range 24.0–82.0) and 66.0  years (39–84), respectively. The Univariate analysis showed that EUS T-stage (Χ 8.321, median follow-up period was 25.0 months in the develop- df 3, p = 0.040), N-st age (Χ 14.879, df 3, p = 0.002), ment cohort (95% CI 23.1–26.9) and 26.0 months (95% and N0 versus N+ (χ 11.325, df 1, p = 0.001) were sig- CI 22.7–29.3) in the validation cohort. Mean survival nificantly associated with OS. When EUS T-stage and times are presented because a 50% mortality rate was not N-stage were entered in model 1, only EUS N-stage was reached in either cohort. The mean OS of the develop- significantly and independently associated with dura- ment cohort was 33.1  months (95% CI 30.1–36.1) and tion of survival [hazard ratio (HR) 1.616–4.707, 95% CI 29.8 months (95% CI 27.1–35.2) in the validation cohort. 0.363–7.190, p = 0.005]. When EUS T-stage and EUS N0 versus N+ were entered in model 2, EUS N0 versus N+ Table 1 Baseline characteristics Frequency (%) Development cohort (n = 117) Validation cohort (n = 139) of patients in development and validation cohorts Male:female 88 (75.2):29 (24.8) 108 (77.7):31 (22.3) Tumour location  Oesophagus 73 (62.4) 76 (54.7)   Upper 0 (0.0) 2 (2.7)   Mid 20 (27.4) 22 (28.9)   Distal 53 (72.6) 52 (68.4)  Junction 44 (37.6) 63 (45.3)   Siewert type I 5 (11.3) 25 (39.7)   Siewert type II 12 (27.3) 18 (28.6)   Siewert type III 27 (61.4) 20 (31.7)  Histology   Adenocarcinoma 98 (83.8) 107 (77.0)   SCC 19 (16.2) 26 (18.7)   HGD 0 (0.0) 3 (2.2)   Neuro-endocrine 0 (0.0) 2 (1.4)   Undifferentiated 0 (0.0) 1 (0.7)  EUS T-stage   T1 18 (15.4) 20 (14.4)   T2 16 (13.7) 18 (12.9)   T3 75 (64.1) 86 (61.9)   T4a 8 (6.8) 13 (9.4)   T4b 0 (0.0) 2 (1.4) EUS N-stage   N0 78 (66.7) 89 (64.0)   N1 23 (19.7) 42 (30.2)   N2 9 (7.6) 7 (5.1)   N3 7 (6.0) 1 (0.7) Treatment  Curative 105 (89.7) 116 (83.5)   NACT 40 (38.1) 44 (37.9)   dCRT 29 (27.6) 39 (33.6)   Surgery alone 32 (30.5) 19 (16.4)   NACRT 1 (0.9) 11 (9.5)   EMR 3 (2.9) 3 (2.6)  Palliative 12 (10.3) 23 (16.5)  Mortality   Alive 84 (71.8) 85 (61.2)   Dead 33 (28.2) 54 (38.8) 1 3 4976 Surgical Endoscopy (2018) 32:4973–4979 Table 2 Survival data of patients in validation cohort derived from univariate analysis EUS variable Mean OS (months) 95% confidence interval Lower Upper T-stage  T1 41.563 38.834 44.291  T2 25.830 20.062 31.598  T3 27.908 24.484 31.332  T4a 16.846 9.842 23.851  T4b 17.500 4.334 30.666 N-stage  N0 31.853 28.735 34.971  N1 25.625 21.246 30.004  N2 16.857 11.873 21.841  N3 17.000 17.000 17.000 Fig. 1 Significant difference in cumulative survival by EUS T-stage  N+ 24.924 20.966 28.882 in Validation cohort (χ 21.031, df 4, p < 0.001). Patients with more advanced EUS T-stage have worse OS Table 3 Results of model 1 multi-variate analysis including EUS was significantly and independently associated with OS T-stage and EUS N-stage in validation cohort (HR 3.105, 95% CI 1.543–6.247, p = 0.001). Variable p value Hazard ratio df 95% confidence interval Univariate analysis in validation cohort Lower Upper EUS T-stage 0.011 4 EUS T-stage (χ 21.031, df 4, p < 0.001) (Fig. 1) and EUS  T2 0.018 12.482 1 1.528 101.957 N0 versus N+ (χ 4.300, df 1, p = 0.038) were significantly  T3 0.016 11.656 1 1.570 86.548 associated with OS. EUS N-stage did not show a statis-  T4a 0.001 30.114 1 3.731 243.079 tically significant association with OS (χ 5.699, df 3,  T4b 0.050 16.270 1 0.994 266.273 p = 0.127). Table 2 shows mean survival data for patients EUS N-stage 0.553 3 classified by EUS T-stage, N-stage, and EUS N0 versus  N1 0.560 1.192 1 0.660 2.154 N+.  N2 0.353 1.653 1 0.572 4.772  N3 0.260 3.176 1 0.425 23.716 Multivariate analysis in validation cohort Table 4 Results of model 2 multi-variate analysis including EUS Again, two Cox regression models were produced in the T-stage and EUS N0 versus N+ in validation cohort validation cohort. EUS T-stage and EUS N-stage were entered in model 1. EUS T-stage was significantly and inde - Variable p value Hazard ratio df 95% confidence interval pendently associated with OS (HR 11.656–30.114, 95% CI 0.994–243.079, p = 0.011). (Table  3) The EPV ratio was Lower Upper 27.0. EUS T-stage 0.012 4 EUS T-stage and EUS N0 versus N+ were entered in  T2 0.020 11.977 1 1.469 97.620 model 2. Again, EUS T-stage was significantly and inde-  T3 0.016 11.714 1 1.579 86.902 pendently associated with OS (HR 11.714–29.631, 95% CI  T4a 0.001 29.631 1 3.674 238.959 1.067–238.959, p = 0.012). (Table  4) The EPV ratio was  T4b 0.045 17.243 1 1.067 278.714 27.0. EUS N0 vs N+ 0.359 1.292 1 0.747 2.235 There was no statistically significant difference in OS between the development (mean OS 33.0 months, 95% CI 30.060–36.076) and validation cohorts (mean OS 29.8 months, 95% CI 27.120–32.513) (χ 1.979, df 1, p = 0.159). 1 3 Surgical Endoscopy (2018) 32:4973–4979 4977 Eec ff t of including site 1 patients in development cohort A post hoc analysis was performed to determine the effect of including patients scanned in site 1 on the development cohort prognostic models. To perform this post hoc analysis, site 1 patients were excluded from the development cohort in an attempt to control comparison with the validation cohort. Seventy patients were originally scanned at site 2. Of these, 53 patients (75.7%) were staged EUS N0; 11 (15.7%) were EUS N1; 3 (4.3%) were N2, and 3 (4.3%) were N3. Both EUS N-stage (HR 2.365–32.757, 95% CI 0.476–223.922, p = 0.005) and EUS N0 versus N+ (HR 3.783, 1.141–12.539, p = 0.030) were independent predictors of OS, in keeping with findings from the original study. Therefore, inclusion of site 1 patients had little effect on the prognostic models of the development cohort. Confidence intervals are wide, Fig. 2 Significant difference in cumulative survival depending on the likely due to the small numbers in N2 and N3 groups. presence of pathological lymph nodes (χ 13.315, df 1, p < 0.001). Patients with positive pathological lymph nodes have worse OS Comparison of patient eligibility for development and validation cohorts (p = 0.038). Relatively small numbers of patients staged EUS N2 and N3 were included in the development cohort A comparison of the proportion of patients who were staged (n = 9 and n = 7, respectively) which could result in dispro- N0 and considered for inclusion during both study periods portionate statistical significance and failure to validate this was made. Post hoc review revealed that 117 of 207 patients variable. Importantly, this study showed that EUS T-stage is (56.5%) from site 2 were staged N0 on PET/CT during the significantly and independently associated with OS, which study recruitment period of the development cohort and 139 supports other studies [14–16]. The study adds evidence to of 317 (43.8%) from site 2 were staged N0 during the study the importance of EUS in the multi-modality approach to recruitment period of the validation cohort. This difference oesophageal cancer staging. was statistically significant (χ 8.049, df 1, p = 0.005). A When patients scanned in site 1 were removed from the significantly higher proportion of patients from site 2 were development cohort, EUS N-stage and EUS N0 versus N+ staged N0 in the development cohort, suggesting that dif- remained independent predictors of OS. This finding sug- ferent proportions of patients were eligible for inclusion, gests that the site 1 PET/CT scanner had little effect on the potentially affecting the results of the prognostic models. developed prognostic model. One reason for the inability to validate the model could be the statistically significant Prognostic significance of pathological lymph nodes difference in proportions of patients staged N0 during both study periods, resulting in fewer patients from site 2 eligi- In total, 74 patients from the validation cohort underwent ble for inclusion in the validation study recruitment period surgical resection. Thirty-nine patients (52.7%) were classi- (those staged N0 on PET/CT). fied pN0 and 35 patients (47.3%) were classified pN+. There An important issue in validation studies is the extent to was a significant difference in OS between pN0 and pN+ which the included cohorts differ. This can result in valida- groups (χ 13.315, df 1, p < 0.001). (Fig. 2) Mean OS for the tion failure unless appropriately adjusted for [17]. Report- pN0 group was 40.091 months (95% CI 36.931–43.251) and ing trends over time have not been assessed in this study, 26.538 (95% CI 22.123–30.953) for the pN+ group. but reporting and context bias have been shown to influence radiologists image interpretation [18]. It is possible that a priori knowledge of key findings from the developed prog- Discussion nostic model may increase the likelihood of equivocal lymph nodes on PET/CT being called positive. This awareness of This validation study failed to replicate the results of the having been studied with consequent impact on behaviour, developed prognostic model. In the validation cohort, EUS the so-called ‘Hawthorne effect’, could have contributed to N-stage and EUS N0 versus N+ did not show prognostic the change in results [19]. However, this is merely a hypoth- significance in multi-variate analysis, although EUS N0 ver - esis and cannot be concluded from these data. In addition, sus N+ was statistically significant in univariate analysis 1 3 4978 Surgical Endoscopy (2018) 32:4973–4979 operators were not fully blinded to the results of the PET/CT, importance of rigorous study design and statistical analy- potentially influencing the interpretation of EUS N-stage. sis cannot be understated, and further research is required The final report of the PET/CT is routinely checked prior to improve prognostic research methodology [24, 25]. In to EUS to ensure that no distant metastases are detected, addition, it is important to document and publish the find - preventing inappropriate EUS examination. ings of all prognostic research, including ‘non-significant’ Prior to the opening of site 2 in 2010, patients were findings, since publication bias is prevalent [26]. scanned at site 1 using a Philips 16 section Gemini GXL This validation study is limited by some factors. Both dedicated PET/CT system (Philips Medical Systems, Cleve- development and validation cohorts represent a relatively land, USA). The two sites used different scanners and proto- heterogeneous cohort of patients; however, this is reflec- cols, and patients were scanned at different activity uptake tive of the demographics of oesophageal cancer patients times. Patients were scanned at 60 min of uptake time in in general. Dissemination of lymph node metastases is site 1 and after 90 min in site 2. Longer uptake times lead dependent on T-stage and to a lesser extent, histological to higher tumour to background avidity and can therefore cell type of the primary tumour [27, 28]. Stratification of increase the conspicuity of lymph node metastases. Sec- these factors was not performed, but most patients pre- ondly, the scanner in site 2 had a time-of-flight (TOF) algo- sented with locally advanced T3 or T4 tumours. The EUS rithm but the site 1 scanner did not. TOF reconstructions operators may have been influenced by results of the PET/ improve signal-to-noise ratio, detection and anatomical CT reports, resulting in inadvertent changes in lymph node localisation of lymph node metastases by allowing more assessment over time. In addition, the inclusion of patients precise measurement of the time difference between detec- from site 1 could not be replicated in the validation cohort tions [20]. Finally, images were acquired for 4 min per bed because our patients no longer attend this site for PET/CT. position in site 1, whereas the acquisition was 3 min per bed Despite the limitations, this validation study has sev- position in site 2. Some improvement in image quality may eral strengths. All patients were managed by an experi- be expected in site 1 with longer acquisition times, provided enced Upper gastrointestinal cancer multi-disciplinary the patient remained still. However, the results of this valida- team covering a population of approximately 1.5 million. tion study assume that longer acquisition did not affect the The site 2 PET/CT scanner and acquisition protocol was models. Standardised PET acquisition protocols have been unchanged in both development and validation cohorts. discussed in the literature [21], but these findings suggest the The study was adequately powered according to a recom- differences may be less influential than previously thought, mended EPV ratio. In addition, a pathologist reviewed the adding generalisability to the results. resected lymph nodes to evaluate and confirm the presence An additional hypothesis for the failure of validation is of pathological lymph node metastases. the accuracy of the staging investigations. Results from our institution have shown suboptimal diagnostic accuracy in the general oesophageal cancer population [22]. Overall accuracy of differentiating negative and positive nodal dis- Conclusion ease on CT, EUS, and PET/CT was 54.5, 55.4, and 57.1%, respectively. Sensitivity and specificity were 39.7 and 77.3% Validation studies are important in prognostic research with CT, 42.6 and 75.0% with EUS and 35.3 and 90.9% with [29]. Lymph node metastases have a significant detrimen- PET/CT. These results were largely attributable to the detec- tal effect on OS. This validation study did not replicate the tion of micro-metastases in 44% of lymph nodes (defined results of the developed prognostic model but the contin- as ≤ 2 mm) evaluated pathologically. The suboptimal accu- ued benefit of EUS in patients staged N0 with PET/CT racy could affect results of the developed prognostic model, was demonstrated. EUS remains a valuable component of obtaining a significant difference in survival between EUS a multi-modality approach to oesophageal cancer staging. N-stage categories by chance alone. Funding No funding was received for this study. The additional sub-group analysis conducted in this study confirms the presence of lymph node metastases as Compliance with ethical standards a major prognostic indicator [2]. There was a highly sig- nificant difference in OS between pN0 and pN+ groups in Disclosures Dr. Kieran Foley, Dr. Patrick Fielding, Professor Wyn patients staged N0 on PET/CT. This finding highlights the Lewis, and Dr. Ashley Roberts have no conflicts of interest and or importance of accurate pre-treatment lymph node staging in financial ties to disclose. oesophageal cancer. Model validation is not commonly performed in prog- Open Access This article is distributed under the terms of the Crea- nostic research. Approximately 34% of models are vali- tive Commons Attribution 4.0 International License (http://creat dated, with 11% undergoing external validation [23]. The iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, 1 3 Surgical Endoscopy (2018) 32:4973–4979 4979 distribution, and reproduction in any medium, provided you give appro- margin involvement following oesophagectomy for cancer and priate credit to the original author(s) and the source, provide a link to the predictive role of endoluminal ultrasonography. Br J Cancer the Creative Commons license, and indicate if changes were made. 107:1925–1931 16. Hiele M, DeLeyn P, Schurmans P, Lerut A, Huys S, Geboes K, Gevers AM, Rutgeerts P (1997) Relation between endoscopic ultrasound findings and outcome of patients with tumors of the References esophagus or esophagogastric junction. Gastrointest Endosc 45:381–386 1. Mukherjee S, Hurt CN, Gwynne S, Bateman A, Gollins S, Rad- 17. Debray TP, Vergouwe Y, Koffijberg H, Nieboer D, Steyerberg EW, hakrishna G, Hawkins M, Canham J, Lewis W, Grabsch HI, Moons KG (2015) A new framework to enhance the interpretation Sharma RA, Wade W, Maggs R, Tranter B, Roberts A, Sebag- of external validation studies of clinical prediction models. J Clin Montefiore D, Maughan T, Griffiths G, Crosby T (2015) NEO- Epidemiol 68:279–289 SCOPE: a randomised phase II study of induction chemotherapy 18. Egglin TK, Feinstein AR (1996) Context bias. A problem in diag- followed by either oxaliplatin/capecitabine or paclitaxel/carbopl- nostic radiology. JAMA 276:1752–1755 atin based chemoradiation as pre-operative regimen for resectable 19. McCambridge J, Witton J, Elbourne DR (2014) Systematic oesophageal adenocarcinoma. BMC Cancer 15:48 review of the Hawthorne effect: new concepts are needed to study 2. Kayani B, Zacharakis E, Ahmed K, Hanna GB (2011) Lymph research participation effects. J Clin Epidemiol 67:267–277 node metastases and prognosis in oesophageal carcinoma—a sys- 20. Surti S (2015) Update on time-of-flight PET imaging. J Nucl Med tematic review. Eur J Surg Oncol 37:747–753 56:98–105 3. Blazeby JM, Farndon JR, Donovan J, Alderson D (2000) A 21. Boellaard R, O’Doherty MJ, Weber WA, Mottaghy FM, Lonsdale prospective longitudinal study examining the quality of life of MN, Stroobants SG, Oyen WJ, Kotzerke J, Hoekstra OS, Pruim J, patients with esophageal carcinoma. Cancer 88:1781–1787 Marsden PK, Tatsch K, Hoekstra CJ, Visser EP, Arends B, Ver- 4. Roberts SA, Kay C (2014) Oesophagus and stomach cancers. In: zijlbergen FJ, Zijlstra JM, Comans EF, Lammertsma AA, Paans Nicholson T (ed) Recommendations for cross-sectional imaging AM, Willemsen AT, Beyer T, Bockisch A, Schaefer-Prokop C, in cancer management. Royal College of Radiologists, London Delbeke D, Baum RP, Chiti A, Krause BJ (2010) FDG PET and 5. National Oesophago-Gastric Cancer Audit (2016) An audit of the PET/CT: EANM procedure guidelines for tumour PET imaging: care received by people with oesophago-gastric cancer in Eng- version 1.0. Eur J Nucl Med Mol Imaging 37:181–200 land and Wales annual report. National Oesophago-Gastric Cancer 22. Foley KG, Christian A, Fielding P, Lewis WG, Roberts SA (2017) Audit, London Accuracy of contemporary oesophageal cancer lymph node 6. van Vliet EP, Heijenbrok-Kal MH, Hunink MG, Kuipers EJ, staging with radiological-pathological correlation. Clin Radiol Siersema PD (2008) Staging investigations for oesophageal can- 72:e691–e697 cer: a meta-analysis. Br J Cancer 98:547–557 23. Mallett S, Royston P, Waters R, Dutton S, Altman DG (2010) 7. Foley KG, Lewis WG, Fielding P, Karran A, Chan D, Blake P, Reporting performance of prognostic models in cancer: a review. Roberts SA (2014) N-staging of oesophageal and junctional car- BMC Med 8:21 cinoma: is there still a role for EUS in patients staged N0 at PET/ 24. Riley RD, Hayden JA, Steyerberg EW, Moons KG, Abrams K, CT? Clin Radiol 69:959–964 Kyzas PA, Malats N, Briggs A, Schroter S, Altman DG, Hem- 8. Kapoor V, McCook BM, Torok FS (2004) An introduction to ingway H, Group P (2013) Prognosis Research Strategy (PRO- PET-CT imaging. Radiographics 24:523–543 GRESS) 2: prognostic factor research. PLoS Med 10:e1001380 9. Sobin LH, Gospodarowicz MK, Wittekind CH (2009) UICC TNM 25. Moons KG, Royston P, Vergouwe Y, Grobbee DE, Altman DG classification of malignant tumours, 7th edn. Wiley, New York (2009) Prognosis and prognostic research: what, why, and how? 10. Kaplan E, Meier P (1958) Nonparametric estimation from incom- BMJ 338:b375 plete observations. J Am Stat Assoc 53:457–481 26. Kyzas PA, Denaxa-Kyza D, Ioannidis JP (2007) Almost all arti- 11. Cox DR (1972) Regression models and life-tables. J R Stat Soc cles on cancer prognostic markers report statistically significant 34:187–220 results. Eur J Cancer 43:2559–2579 12. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR (1996) 27. Siewert JR, Stein HJ, Feith M, Bruecher BL, Bartels H, Fink U A simulation study of the number of events per variable in logistic (2001) Histologic tumor type is an independent prognostic param- regression analysis. J Clin Epidemiol 49:1373–1379 eter in esophageal cancer: lessons from more than 1,000 consecu- 13. Collins GS, Reitsma JB, Altman DG, Moons KG (2015) Transpar- tive resections at a single center in the Western world. Ann Surg ent reporting of a multivariable prediction model for individual 234:360–369 prognosis or diagnosis (TRIPOD): the TRIPOD statement. BMJ 28. Rice TW, Zuccaro G Jr, Adelstein DJ, Rybicki LA, Blackstone 350:g7594 EH, Goldblum JR (1998) Esophageal carcinoma: depth of tumor 14. Paraf F, Flejou JF, Pignon JP, Fekete F, Potet F (1995) Surgi- invasion is predictive of regional lymph node status. Ann Thorac cal pathology of adenocarcinoma arising in Barrett’s esophagus. Surg 65:787–792 Analysis of 67 cases. Am J Surg Pathol 19:183–191 29. Altman DG, Royston P (2000) What do we mean by validating a 15. Reid TD, Chan DS, Roberts SA, Crosby TD, Williams GT, Lewis prognostic model? Stat Med 19:453–473 WG (2012) Prognostic significance of circumferential resection 1 3

Journal

Surgical EndoscopySpringer Journals

Published: Jun 4, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off