TY - JOUR
AU - Andersson, R
AB - Abstract Background The size of pancreatic ductal adenocarcinoma (PDAC) at diagnosis is an indicator of outcome. Previous studies have focused mostly on patients with resectable disease. The aim of this study was to investigate the relationship between tumour size and risk of metastasis and death in a large PDAC cohort, including all stages. Methods Patients diagnosed with PDAC between 1988 and 2013 were identified from the Surveillance, Epidemiology, and End Results (SEER) database. Tumour size was defined as the maximum dimension of the tumour as provided by the registry. Metastatic spread was assessed, and survival was calculated according to size of the primary tumour using the Kaplan–Meier method. Cox proportional regression modelling was used to adjust for known confounders. Results Some 58 728 patients were included. There were 187 patients (0·3 per cent) with a tumour size of 0·5 cm or less, in whom the rate of distant metastasis was 30·6 per cent. The probability of tumour dissemination was associated with tumour size at the time of diagnosis. The association between survival and tumour size was linear for patients with localized tumours, but stochastic in patients with regional and distant stages. In patients with resected tumours, increasing tumour size was associated with worse tumour-specific survival, whereas size was not associated with survival in patients with unresected tumours. In the adjusted Cox regression analysis, the death rate increased by 4·1 per cent for each additional 1-cm increase in tumour size. Conclusion Pancreatic cancer has a high metastatic capacity even in small tumours. The prognostic impact of tumour size is restricted to patients with localized disease. Introduction Pancreatic ductal adenocarcinoma (PDAC) is associated with an overall 5-year survival rate of less than 7 per cent1. Multiple factors are responsible for the poor prognosis, including late presentation, aggressive tumour biology and the lack of effective systemic therapies2. Although surgery provides a chance of cure, the 5-year survival rate after surgery is only 20 per cent3. PDAC is believed to arise from non-invasive precursor lesions including pancreatic intraepithelial neoplasia, intraductal papillary mucinous neoplasm and mucinous cystic neoplasm4. Progression of precursor lesions to invasive PDAC is likely to be a slow process that develops over years, or even decades5. Some studies6,7 have suggested that, once cancer has developed, progression from an early to a more advanced stage can occur quickly. Small tumours may already have distant metastases at the time of diagnosis8. Traditionally, tumour size has been viewed as an important prognostic factor. According to the TNM classification system, T1 tumours are defined as lesions with a size of 2 cm or less confined to the pancreas9,10. These small tumours are considered to be at an early stage of the disease and have a favourable prognosis11. The development of metastasis is mostly considered as a late event in the progression of PDAC, occurring mainly in larger tumours5,8,12,13. Several studies6,14–16, however, have suggested that even small tumours are associated with disseminated disease, sometimes even before PDAC has reached the detection limit. The aim of this study was to investigate the relationship between primary tumour size and metastatic rates and survival in patients with PDAC. It was hypothesized that tumours no larger than 0·5 cm in diameter are already associated with locoregional and distant metastases. Methods The National Cancer Institute Surveillance, Epidemiology, and End Results (SEER) database is a national programme of 18 regional or statewide cancer registries in the USA. Data were obtained from all US cancer registries participating in the SEER programme using SEER*Stat version 8.3.217. The Ethics Committee for Clinical Research at Lund University, Sweden, approved the study protocol (2016/100). The study followed the STROBE guidelines18. Patients and study design The study group included all patients with pancreatic cancer registered in the SEER database between 1988 and 2013. Patients were identified on the basis of the ICD for Oncology, third edition (ICD-O-3) for tumours of the exocrine pancreas: C25.0, C25.1, C25.2, C25.3, C25.7, C25.8 and C25.9. Only patients with microscopically confirmed infiltrating pancreatic adenocarcinoma or PDAC (ICD-O-3 histology codes 8140 and 8500 respectively) were selected. Patients with non-invasive tumours, mucinous cystic neoplasms or histological variants, such as adenosquamous carcinoma, colloid carcinoma or hepatoid adenocarcinoma, were excluded. Tumour size was measured as the maximum length of the tumour based on the pathological, operative or radiological report, in this order of priority. Patients were excluded when tumour size was not defined or the value recorded was greater than 20 cm. Patients with incomplete follow-up were also removed. Disease-specific survival was calculated from diagnosis to date of death, last date known to be alive, or until last follow-up (November 2015). Individuals who died from causes other than pancreatic cancer were censored. To ensure a coherent cancer staging classification across the study period, the SEER historical stage A was used, which provides a consistent definition over time. The AJCC staging system, which is used more widely in clinical settings, was not accessible for many of the annual data sets analysed. The SEER historical stages were outlined as localized (limited to the pancreas; AJCC IA or IB), regional (tumour invading adjacent structures or spread to regional lymph nodes; AJCC IIA, IIB or III) or distant (presence of distant metastasis; AJCC IV). Statistical analysis Data were analysed using Stata® MP 14.1 software (StataCorp, College Station, Texas, USA). Survival curves were calculated using the Kaplan–Meier method. The Cox proportional hazards model was used to adjust for confounding variables. Proportional hazards assumptions were checked, giving more weight to the graphical tests than the statistical ones because of the large sample size. A Poisson model was used to verify the results, with survival time as an offset. Factors of interest were: age, sex, tumour stage, histological grade, tumour location, surgical resection, radiation therapy, chemotherapy and year of diagnosis. Variable selection was determined by the literature and background medical knowledge; collinearity was checked and variables removed if collinearity was deemed problematic, resulting in the main effect model with no interaction terms. Hazard ratios are presented with 95 per cent confidence intervals. Missing values were imputed using the multiple imputation with chained equations technique, as described by White and colleagues19. The imputation method was predictive mean matching. The number of iterations for each chain was ten, as was the number of imputed data sets. Results In total, 177 115 patients with pancreatic cancer were registered in the SEER database. Some 118 387 patients did not fulfil the inclusion criteria and were excluded. Histological or cytological confirmation of the tumour was not available in 70 035 patients. Tumour size was not defined in 35 571 patients and data on outcome or follow-up was not available for 12 781. The study group consisted of 58 728 patients (Fig. 1). Fig. 1 Open in new tabDownload slide Flow diagram for the study Some 8·1 per cent of the patients had localized disease only, 40·2 per cent had regional disease, 49·6 per cent had distant disease and in 2·1 per cent of patients the disease stage was unknown (Table 1). Table 1 Cohort characteristics . No. of patients* (n = 58 728) . Age (years)† 67·0(11·5) Sex ratio (F : M) 29 003 : 29 725 Tumour size (cm) ≤ 0·5 187 (0·3) 0·6–1·0 430 (0·7) 1·1–1·5 1349 (2·3)  1·6–2·0 3791 (6·5)  > 2·0 52 971 (90·2)  Histological grade Well differentiated 3142 (5·4)  Moderately differentiated 12 409 (21·1)  Poorly differentiated 11 484 (19·6)  Anaplastic 470 (0·8) Unknown 31 223 (53·2)  Tumour location Head 34 792 (59·2)  Body 7564 (12·9) Tail 7470 (12·7) Pancreatic duct 356 (0·6) Other specified parts of pancreas 758 (1·3) Overlapping 4752 (8·1)  Pancreas, NOS 3036 (5·2)  Stage Localized 4777 (8·1)  Regional 23 579 (40·2)   Distant 29 128 (49·6)   Unknown 1244 (2·1)  Surgical resection No 43 182 (73·5)  Yes 15 398 (26·2)  Unknown 148 (0·3) Radiation therapy No 43 011 (73·2)  Yes 15 717 (26·8)  Chemotherapy No‡ 26 592 (45·3)  Yes 32 136 (54·7)  Time period 1988–2006 27 601 (47·0)  2007–2013 31 127 (53·0)  . No. of patients* (n = 58 728) . Age (years)† 67·0(11·5) Sex ratio (F : M) 29 003 : 29 725 Tumour size (cm) ≤ 0·5 187 (0·3) 0·6–1·0 430 (0·7) 1·1–1·5 1349 (2·3)  1·6–2·0 3791 (6·5)  > 2·0 52 971 (90·2)  Histological grade Well differentiated 3142 (5·4)  Moderately differentiated 12 409 (21·1)  Poorly differentiated 11 484 (19·6)  Anaplastic 470 (0·8) Unknown 31 223 (53·2)  Tumour location Head 34 792 (59·2)  Body 7564 (12·9) Tail 7470 (12·7) Pancreatic duct 356 (0·6) Other specified parts of pancreas 758 (1·3) Overlapping 4752 (8·1)  Pancreas, NOS 3036 (5·2)  Stage Localized 4777 (8·1)  Regional 23 579 (40·2)   Distant 29 128 (49·6)   Unknown 1244 (2·1)  Surgical resection No 43 182 (73·5)  Yes 15 398 (26·2)  Unknown 148 (0·3) Radiation therapy No 43 011 (73·2)  Yes 15 717 (26·8)  Chemotherapy No‡ 26 592 (45·3)  Yes 32 136 (54·7)  Time period 1988–2006 27 601 (47·0)  2007–2013 31 127 (53·0)  * With percentages in parentheses unless indicated otherwise; † values are mean(s.d.). ‡ No evidence of chemotherapy found in the medical records examined. NOS, not otherwise specified. Open in new tab Table 1 Cohort characteristics . No. of patients* (n = 58 728) . Age (years)† 67·0(11·5) Sex ratio (F : M) 29 003 : 29 725 Tumour size (cm) ≤ 0·5 187 (0·3) 0·6–1·0 430 (0·7) 1·1–1·5 1349 (2·3)  1·6–2·0 3791 (6·5)  > 2·0 52 971 (90·2)  Histological grade Well differentiated 3142 (5·4)  Moderately differentiated 12 409 (21·1)  Poorly differentiated 11 484 (19·6)  Anaplastic 470 (0·8) Unknown 31 223 (53·2)  Tumour location Head 34 792 (59·2)  Body 7564 (12·9) Tail 7470 (12·7) Pancreatic duct 356 (0·6) Other specified parts of pancreas 758 (1·3) Overlapping 4752 (8·1)  Pancreas, NOS 3036 (5·2)  Stage Localized 4777 (8·1)  Regional 23 579 (40·2)   Distant 29 128 (49·6)   Unknown 1244 (2·1)  Surgical resection No 43 182 (73·5)  Yes 15 398 (26·2)  Unknown 148 (0·3) Radiation therapy No 43 011 (73·2)  Yes 15 717 (26·8)  Chemotherapy No‡ 26 592 (45·3)  Yes 32 136 (54·7)  Time period 1988–2006 27 601 (47·0)  2007–2013 31 127 (53·0)  . No. of patients* (n = 58 728) . Age (years)† 67·0(11·5) Sex ratio (F : M) 29 003 : 29 725 Tumour size (cm) ≤ 0·5 187 (0·3) 0·6–1·0 430 (0·7) 1·1–1·5 1349 (2·3)  1·6–2·0 3791 (6·5)  > 2·0 52 971 (90·2)  Histological grade Well differentiated 3142 (5·4)  Moderately differentiated 12 409 (21·1)  Poorly differentiated 11 484 (19·6)  Anaplastic 470 (0·8) Unknown 31 223 (53·2)  Tumour location Head 34 792 (59·2)  Body 7564 (12·9) Tail 7470 (12·7) Pancreatic duct 356 (0·6) Other specified parts of pancreas 758 (1·3) Overlapping 4752 (8·1)  Pancreas, NOS 3036 (5·2)  Stage Localized 4777 (8·1)  Regional 23 579 (40·2)   Distant 29 128 (49·6)   Unknown 1244 (2·1)  Surgical resection No 43 182 (73·5)  Yes 15 398 (26·2)  Unknown 148 (0·3) Radiation therapy No 43 011 (73·2)  Yes 15 717 (26·8)  Chemotherapy No‡ 26 592 (45·3)  Yes 32 136 (54·7)  Time period 1988–2006 27 601 (47·0)  2007–2013 31 127 (53·0)  * With percentages in parentheses unless indicated otherwise; † values are mean(s.d.). ‡ No evidence of chemotherapy found in the medical records examined. NOS, not otherwise specified. Open in new tab Median tumour diameter was 3·9 (range 0·1–20·0) cm (Fig. S1, supporting information). Only 0·3 per cent of the tumours had a diameter of 0·5 cm or less and 90·2 per cent of the tumours were larger than 2 cm (Table 1). Median follow-up time was 5 (range 0–301) months. Association between tumour size and distant metastasis The risk of distant metastasis at diagnosis increased in a non-linear fashion with increasing primary tumour burden (Fig. 2). Tumour stage distribution according to tumour size is shown in Fig. S2 (supporting information). When the tumour size was no more than 0·5 cm, the rate of distant metastasis was 30·6 per cent, increasing to 73·9 per cent when the tumour size was 10 cm. Fig. 2 Open in new tabDownload slide Fractional polynomial plot illustrating rate of distant metastasis in relation to tumour size at the time of diagnosis Survival according to tumour size and stage The 5-year survival rate for the study group was 5·4 per cent. For tumours no larger than 0·5 cm the 5-year survival rate was 23·1 per cent. Survival decreased with increasing tumour size (Table 2). Among 4777 patients with local disease only, the 5-year survival rate was 16·0 per cent; it was 57 per cent for patients with tumours of 0·5 cm or less in diameter, decreasing to 12·0 per cent for those with tumours larger than 2 cm. In patients with regional disease or distant metastases, survival was similar across tumour size categories. The relationship between tumour size and survival, stratified by stage, is shown in Fig. 3 and Fig. S3 (supporting information). Fig. 3 Open in new tabDownload slide Kaplan–Meier curves illustrating survival by tumour size (cm) and stage of disease: a localized, b regional, c distant and d all stages Table 2 Tumour size and survival according to disease stage Tumour size (cm) . No. of patients . Survival (%)* . 3 years . 5 years  . All stages 58 728 (100) 8·4 5·4 ≤ 0·5 187 (0·3) 26·7 23·1 0·6–1·0 430 (0·7) 26·6 20·4 1·1–1·5 1349 (2·3) 24·9 17·3 1·6–2·0 3791 (6·5) 16·9 11·3 > 2·0 52 971 (90·2) 7·1 4·4 Localized disease 4777 (100) 20·3 16·0 ≤ 0·5 66 (1·4) 63 57 0·6–1·0 128 (2·7) 49·3 39·7 1·1–1·5 265 (5·5) 39·7 27·5 1·6–2·0 570 (11·9) 31·2 26·8 > 2·0 3748 (78·5) 15·4 12·0 Regional disease 23 579 (100) 13·5 8·2 ≤ 0·5 59 (0·3) 14 7 0·6–1·0 152 (0·6) 30·3 22·7 1·1–1·5 650 (2·8) 29·9 21·2 1·6–2·0 1765 (7·5) 23·1 14·5 > 2·0 20 953 (88·9) 11·9 7·1 Distant metastases 29 128 (100) 2·2 1·1 ≤ 0·5 55 (0·2) 0 0 0·6–1·0 148 (0·5) 3·2 2·1 1·1–1·5 411 (1·4) 7·4 3·2 1·6–2·0 1364 (4·7) 3·4 1·5 > 2·0 27 150 (93·2) 2·0 1·1 Tumour size (cm) . No. of patients . Survival (%)* . 3 years . 5 years  . All stages 58 728 (100) 8·4 5·4 ≤ 0·5 187 (0·3) 26·7 23·1 0·6–1·0 430 (0·7) 26·6 20·4 1·1–1·5 1349 (2·3) 24·9 17·3 1·6–2·0 3791 (6·5) 16·9 11·3 > 2·0 52 971 (90·2) 7·1 4·4 Localized disease 4777 (100) 20·3 16·0 ≤ 0·5 66 (1·4) 63 57 0·6–1·0 128 (2·7) 49·3 39·7 1·1–1·5 265 (5·5) 39·7 27·5 1·6–2·0 570 (11·9) 31·2 26·8 > 2·0 3748 (78·5) 15·4 12·0 Regional disease 23 579 (100) 13·5 8·2 ≤ 0·5 59 (0·3) 14 7 0·6–1·0 152 (0·6) 30·3 22·7 1·1–1·5 650 (2·8) 29·9 21·2 1·6–2·0 1765 (7·5) 23·1 14·5 > 2·0 20 953 (88·9) 11·9 7·1 Distant metastases 29 128 (100) 2·2 1·1 ≤ 0·5 55 (0·2) 0 0 0·6–1·0 148 (0·5) 3·2 2·1 1·1–1·5 411 (1·4) 7·4 3·2 1·6–2·0 1364 (4·7) 3·4 1·5 > 2·0 27 150 (93·2) 2·0 1·1 Values in parentheses are percentages. * Calculated by Kaplan–Meier analysis. Open in new tab Table 2 Tumour size and survival according to disease stage Tumour size (cm) . No. of patients . Survival (%)* . 3 years . 5 years  . All stages 58 728 (100) 8·4 5·4 ≤ 0·5 187 (0·3) 26·7 23·1 0·6–1·0 430 (0·7) 26·6 20·4 1·1–1·5 1349 (2·3) 24·9 17·3 1·6–2·0 3791 (6·5) 16·9 11·3 > 2·0 52 971 (90·2) 7·1 4·4 Localized disease 4777 (100) 20·3 16·0 ≤ 0·5 66 (1·4) 63 57 0·6–1·0 128 (2·7) 49·3 39·7 1·1–1·5 265 (5·5) 39·7 27·5 1·6–2·0 570 (11·9) 31·2 26·8 > 2·0 3748 (78·5) 15·4 12·0 Regional disease 23 579 (100) 13·5 8·2 ≤ 0·5 59 (0·3) 14 7 0·6–1·0 152 (0·6) 30·3 22·7 1·1–1·5 650 (2·8) 29·9 21·2 1·6–2·0 1765 (7·5) 23·1 14·5 > 2·0 20 953 (88·9) 11·9 7·1 Distant metastases 29 128 (100) 2·2 1·1 ≤ 0·5 55 (0·2) 0 0 0·6–1·0 148 (0·5) 3·2 2·1 1·1–1·5 411 (1·4) 7·4 3·2 1·6–2·0 1364 (4·7) 3·4 1·5 > 2·0 27 150 (93·2) 2·0 1·1 Tumour size (cm) . No. of patients . Survival (%)* . 3 years . 5 years  . All stages 58 728 (100) 8·4 5·4 ≤ 0·5 187 (0·3) 26·7 23·1 0·6–1·0 430 (0·7) 26·6 20·4 1·1–1·5 1349 (2·3) 24·9 17·3 1·6–2·0 3791 (6·5) 16·9 11·3 > 2·0 52 971 (90·2) 7·1 4·4 Localized disease 4777 (100) 20·3 16·0 ≤ 0·5 66 (1·4) 63 57 0·6–1·0 128 (2·7) 49·3 39·7 1·1–1·5 265 (5·5) 39·7 27·5 1·6–2·0 570 (11·9) 31·2 26·8 > 2·0 3748 (78·5) 15·4 12·0 Regional disease 23 579 (100) 13·5 8·2 ≤ 0·5 59 (0·3) 14 7 0·6–1·0 152 (0·6) 30·3 22·7 1·1–1·5 650 (2·8) 29·9 21·2 1·6–2·0 1765 (7·5) 23·1 14·5 > 2·0 20 953 (88·9) 11·9 7·1 Distant metastases 29 128 (100) 2·2 1·1 ≤ 0·5 55 (0·2) 0 0 0·6–1·0 148 (0·5) 3·2 2·1 1·1–1·5 411 (1·4) 7·4 3·2 1·6–2·0 1364 (4·7) 3·4 1·5 > 2·0 27 150 (93·2) 2·0 1·1 Values in parentheses are percentages. * Calculated by Kaplan–Meier analysis. Open in new tab Survival in patients who underwent surgery Some 15 398 patients underwent surgical resection. The 5-year survival rate was 16·1 per cent after surgery compared with 1·2 per cent for patients who were not operated on. When surgery was performed, patients with tumours no larger than 0·5 cm had better survival than those with larger tumours (Fig. 4). In contrast, in the absence of tumour resection, overall 5-year survival was not associated with tumour size. Fig. 4 Open in new tabDownload slide Kaplan–Meier survival curves showing survival by tumour size (cm) a with or b without surgical resection Predictors of survival All factors considered in the multivariable Cox regression analysis, including age, sex, tumour size, grade, stage, tumour location, surgical resection, radiation therapy and chemotherapy, were associated with cancer-specific survival. For each additional 1-cm increase in tumour size, the rate of death increased by 4·1 per cent (Table 3). Table 3 Predictors of survival identified by multivariable Cox regression analysis . Hazard ratio . P . Age (per year) 1·01 (1·01, 1·01) < 0·001 Male sex 1·03 (1·01, 1·05)   0·001 Size (per cm) 1·04 (1·04, 1·05) < 0·001 Grade Well differentiated 1·00 (reference) Moderately differentiated 1·21 (1·16, 1·26) < 0·001 Poorly differentiated 1·52 (1·45, 1·59) < 0·001 Anaplastic 1·49 (1·32, 1·68) < 0·001 Stage Localized disease 1·00 (reference) Regional disease 1·32 (1·27, 1·37) < 0·001 Distant metastases 2·07 (1·99, 2·15) < 0·001 Location (head of pancreas versus other sites) 0·96 (0·94, 0·98) < 0·001 Surgery 0·43 (0·42, 0·44) < 0·001 Radiation therapy 0·90 (0·88, 0·93) < 0·001 Chemotherapy 0·58 (0·57, 0·59) < 0·001 Time interval (2007–2013 versus 1988–2006) 0·87 (0·85, 0·88) < 0·001 . Hazard ratio . P . Age (per year) 1·01 (1·01, 1·01) < 0·001 Male sex 1·03 (1·01, 1·05)   0·001 Size (per cm) 1·04 (1·04, 1·05) < 0·001 Grade Well differentiated 1·00 (reference) Moderately differentiated 1·21 (1·16, 1·26) < 0·001 Poorly differentiated 1·52 (1·45, 1·59) < 0·001 Anaplastic 1·49 (1·32, 1·68) < 0·001 Stage Localized disease 1·00 (reference) Regional disease 1·32 (1·27, 1·37) < 0·001 Distant metastases 2·07 (1·99, 2·15) < 0·001 Location (head of pancreas versus other sites) 0·96 (0·94, 0·98) < 0·001 Surgery 0·43 (0·42, 0·44) < 0·001 Radiation therapy 0·90 (0·88, 0·93) < 0·001 Chemotherapy 0·58 (0·57, 0·59) < 0·001 Time interval (2007–2013 versus 1988–2006) 0·87 (0·85, 0·88) < 0·001 Values in parentheses are 95 per cent confidence intervals. Open in new tab Table 3 Predictors of survival identified by multivariable Cox regression analysis . Hazard ratio . P . Age (per year) 1·01 (1·01, 1·01) < 0·001 Male sex 1·03 (1·01, 1·05)   0·001 Size (per cm) 1·04 (1·04, 1·05) < 0·001 Grade Well differentiated 1·00 (reference) Moderately differentiated 1·21 (1·16, 1·26) < 0·001 Poorly differentiated 1·52 (1·45, 1·59) < 0·001 Anaplastic 1·49 (1·32, 1·68) < 0·001 Stage Localized disease 1·00 (reference) Regional disease 1·32 (1·27, 1·37) < 0·001 Distant metastases 2·07 (1·99, 2·15) < 0·001 Location (head of pancreas versus other sites) 0·96 (0·94, 0·98) < 0·001 Surgery 0·43 (0·42, 0·44) < 0·001 Radiation therapy 0·90 (0·88, 0·93) < 0·001 Chemotherapy 0·58 (0·57, 0·59) < 0·001 Time interval (2007–2013 versus 1988–2006) 0·87 (0·85, 0·88) < 0·001 . Hazard ratio . P . Age (per year) 1·01 (1·01, 1·01) < 0·001 Male sex 1·03 (1·01, 1·05)   0·001 Size (per cm) 1·04 (1·04, 1·05) < 0·001 Grade Well differentiated 1·00 (reference) Moderately differentiated 1·21 (1·16, 1·26) < 0·001 Poorly differentiated 1·52 (1·45, 1·59) < 0·001 Anaplastic 1·49 (1·32, 1·68) < 0·001 Stage Localized disease 1·00 (reference) Regional disease 1·32 (1·27, 1·37) < 0·001 Distant metastases 2·07 (1·99, 2·15) < 0·001 Location (head of pancreas versus other sites) 0·96 (0·94, 0·98) < 0·001 Surgery 0·43 (0·42, 0·44) < 0·001 Radiation therapy 0·90 (0·88, 0·93) < 0·001 Chemotherapy 0·58 (0·57, 0·59) < 0·001 Time interval (2007–2013 versus 1988–2006) 0·87 (0·85, 0·88) < 0·001 Values in parentheses are 95 per cent confidence intervals. Open in new tab Subgroup analyses were undertaken to determine whether the prognostic impact of tumour size was consistent across stage and treatment categories, adjusting for the same co-variables as the main Cox regression model. Tumour size was associated with an increase in the hazard ratio for death in patients with localized disease. However, no such association could be found for patients with regional disease or distant metastases (Table 4). Increasing tumour size was associated with worse survival in patients with resected tumours, but not in those with unresected disease (Table 5). Table 4 Subgroup analyses of survival according to size and stage Tumour size (cm) . Hazard ratio* . P  . Localized disease ≤ 0·5 1·00 (reference) 0·6–1·0 1·57 (0·94, 2·62) 0·085  1·1–1·5 2·10 (1·31, 3·37) 0·002  1·6–2·0 2·31 (1·46, 3·65) < 0·001  > 2·0 3·38 (2·16, 5·30) < 0·001  Regional disease ≤ 0·5 1·00 (reference) 0·6–1·0 0·74 (0·52, 1·05) 0·088  1·1–1·5 0·73 (0·54, 1·00) 0·047  1·6–2·0 0·88 (0·65, 1·18) 0·398  > 2·0 1·09 (0·82, 1·46) 0·550  Distant metastases ≤ 0·5 1·00 (reference) 0·6–1·0 1·01 (0·71, 1·44) 0·964  1·1–1·5 0·77 (0·56, 1·06) 0·113  1·6–2·0 0·87 (0·64, 1·18) 0·364  > 2·0 0·97 (0·72, 1·31) 0·840  Tumour size (cm) . Hazard ratio* . P  . Localized disease ≤ 0·5 1·00 (reference) 0·6–1·0 1·57 (0·94, 2·62) 0·085  1·1–1·5 2·10 (1·31, 3·37) 0·002  1·6–2·0 2·31 (1·46, 3·65) < 0·001  > 2·0 3·38 (2·16, 5·30) < 0·001  Regional disease ≤ 0·5 1·00 (reference) 0·6–1·0 0·74 (0·52, 1·05) 0·088  1·1–1·5 0·73 (0·54, 1·00) 0·047  1·6–2·0 0·88 (0·65, 1·18) 0·398  > 2·0 1·09 (0·82, 1·46) 0·550  Distant metastases ≤ 0·5 1·00 (reference) 0·6–1·0 1·01 (0·71, 1·44) 0·964  1·1–1·5 0·77 (0·56, 1·06) 0·113  1·6–2·0 0·87 (0·64, 1·18) 0·364  > 2·0 0·97 (0·72, 1·31) 0·840  Values in parentheses are 95 per cent confidence intervals. * Multivariable Cox regression analysis, adjusted for age, male sex, grade, tumour location, surgery, radiation therapy, chemotherapy and time interval. Open in new tab Table 4 Subgroup analyses of survival according to size and stage Tumour size (cm) . Hazard ratio* . P  . Localized disease ≤ 0·5 1·00 (reference) 0·6–1·0 1·57 (0·94, 2·62) 0·085  1·1–1·5 2·10 (1·31, 3·37) 0·002  1·6–2·0 2·31 (1·46, 3·65) < 0·001  > 2·0 3·38 (2·16, 5·30) < 0·001  Regional disease ≤ 0·5 1·00 (reference) 0·6–1·0 0·74 (0·52, 1·05) 0·088  1·1–1·5 0·73 (0·54, 1·00) 0·047  1·6–2·0 0·88 (0·65, 1·18) 0·398  > 2·0 1·09 (0·82, 1·46) 0·550  Distant metastases ≤ 0·5 1·00 (reference) 0·6–1·0 1·01 (0·71, 1·44) 0·964  1·1–1·5 0·77 (0·56, 1·06) 0·113  1·6–2·0 0·87 (0·64, 1·18) 0·364  > 2·0 0·97 (0·72, 1·31) 0·840  Tumour size (cm) . Hazard ratio* . P  . Localized disease ≤ 0·5 1·00 (reference) 0·6–1·0 1·57 (0·94, 2·62) 0·085  1·1–1·5 2·10 (1·31, 3·37) 0·002  1·6–2·0 2·31 (1·46, 3·65) < 0·001  > 2·0 3·38 (2·16, 5·30) < 0·001  Regional disease ≤ 0·5 1·00 (reference) 0·6–1·0 0·74 (0·52, 1·05) 0·088  1·1–1·5 0·73 (0·54, 1·00) 0·047  1·6–2·0 0·88 (0·65, 1·18) 0·398  > 2·0 1·09 (0·82, 1·46) 0·550  Distant metastases ≤ 0·5 1·00 (reference) 0·6–1·0 1·01 (0·71, 1·44) 0·964  1·1–1·5 0·77 (0·56, 1·06) 0·113  1·6–2·0 0·87 (0·64, 1·18) 0·364  > 2·0 0·97 (0·72, 1·31) 0·840  Values in parentheses are 95 per cent confidence intervals. * Multivariable Cox regression analysis, adjusted for age, male sex, grade, tumour location, surgery, radiation therapy, chemotherapy and time interval. Open in new tab Table 5 Subgroup analyses of survival by size, with or without surgical resection Tumour size (cm) . Hazard ratio* . P . Surgical resection ≤ 0·5 1·00 (reference) 0·6–1·0 1·25 (0·83, 1·87) 0·283  1·1–1·5 1·47 (1·01, 2·14) 0·044  1·6–2·0 1·73 (1·20, 2·50) 0·003  > 2·0 2·33 (1·62, 3·36) < 0·001  No surgical resection ≤ 0·5 1·00 (reference) 0·6–1·0 1·04 (0·79, 1·37) 0·762  1·1–1·5 0·82 (0·65, 1·05) 0·116  1·6–2·0 0·93 (0·74, 1·16) 0·504  > 2·0 1·05 (0·84, 1·31) 0·679  Tumour size (cm) . Hazard ratio* . P . Surgical resection ≤ 0·5 1·00 (reference) 0·6–1·0 1·25 (0·83, 1·87) 0·283  1·1–1·5 1·47 (1·01, 2·14) 0·044  1·6–2·0 1·73 (1·20, 2·50) 0·003  > 2·0 2·33 (1·62, 3·36) < 0·001  No surgical resection ≤ 0·5 1·00 (reference) 0·6–1·0 1·04 (0·79, 1·37) 0·762  1·1–1·5 0·82 (0·65, 1·05) 0·116  1·6–2·0 0·93 (0·74, 1·16) 0·504  > 2·0 1·05 (0·84, 1·31) 0·679  Values in parentheses are 95 per cent confidence intervals. * Multivariable Cox regression analysis, adjusted for age, male sex, grade, tumour location, stage, radiation therapy, chemotherapy and time interval. Open in new tab Table 5 Subgroup analyses of survival by size, with or without surgical resection Tumour size (cm) . Hazard ratio* . P . Surgical resection ≤ 0·5 1·00 (reference) 0·6–1·0 1·25 (0·83, 1·87) 0·283  1·1–1·5 1·47 (1·01, 2·14) 0·044  1·6–2·0 1·73 (1·20, 2·50) 0·003  > 2·0 2·33 (1·62, 3·36) < 0·001  No surgical resection ≤ 0·5 1·00 (reference) 0·6–1·0 1·04 (0·79, 1·37) 0·762  1·1–1·5 0·82 (0·65, 1·05) 0·116  1·6–2·0 0·93 (0·74, 1·16) 0·504  > 2·0 1·05 (0·84, 1·31) 0·679  Tumour size (cm) . Hazard ratio* . P . Surgical resection ≤ 0·5 1·00 (reference) 0·6–1·0 1·25 (0·83, 1·87) 0·283  1·1–1·5 1·47 (1·01, 2·14) 0·044  1·6–2·0 1·73 (1·20, 2·50) 0·003  > 2·0 2·33 (1·62, 3·36) < 0·001  No surgical resection ≤ 0·5 1·00 (reference) 0·6–1·0 1·04 (0·79, 1·37) 0·762  1·1–1·5 0·82 (0·65, 1·05) 0·116  1·6–2·0 0·93 (0·74, 1·16) 0·504  > 2·0 1·05 (0·84, 1·31) 0·679  Values in parentheses are 95 per cent confidence intervals. * Multivariable Cox regression analysis, adjusted for age, male sex, grade, tumour location, stage, radiation therapy, chemotherapy and time interval. Open in new tab Discussion In the present study of patients with histologically or cytologically confirmed PDAC, distant metastases were present in almost one-third of patients with tumours of 0·5 cm or less in diameter. This finding may influence staging and treatment of the disease. Identification of new biomarkers for the detection of early disease and targets for treatment is needed. The present study also revealed a stage-dependent relationship between tumour size and survival. The association between small tumour size and prolonged survival was confirmed only in the subgroup of patients with localized cancer. Once the cancer had disseminated, tumour size was no longer an important predictive factor for long-term survival. In some patients, small tumours with regional or distant spread were even associated with a worse prognosis than larger tumours. This suggests that PDAC is a systemic disease already in the early stages of cancer development. PDAC has a high metastatic potential regardless of tumour size. An important observation from the present data is that patients with PDAC had a poor prognosis, even when the disease seemed to be limited to the pancreas. The poor survival in patients diagnosed with localized tumours is generally attributed to early vascular dissemination and metastasis in combination with the insensitivity to adjuvant therapy20,21. In a previous study22, the 5-year survival rate for patients with small localized lesions (less than 2 cm in diameter) was estimated to reach 40 per cent at most. The present results indicate that the survival rate is even lower at a population level. The most common cause of death for patients with resected pancreatic tumours is systemic recurrence rather than local disease23,24, which supports the hypothesis that PDAC is a systemic disease at the time of diagnosis in the majority of patients. Yet, not all PDACs show such aggressiveness at an early stage. PDAC is a heterogeneous disease, and it may be speculated that there are biological differences between tumours that metastasize early and those that behave in a more indolent fashion. Integrated analysis of genomic, epigenomic and transcriptomic data has revealed distinct molecular subtypes of PDAC with different histopathological characteristics and prognosis25. The initiation and progression of PDAC probably take place over years and involve the accumulation of multiple genetic or epigenetic alterations within the cells, resulting in uncontrollable cell expansion with subsequent acquisition of migratory properties and metastatic disease5,26–27. According to a previous study28 that predicted the metastatic capability of PDAC using a mathematical model, the probability of tumour dissemination is dependent on the size of the primary tumour at the time of diagnosis. The present clinical data confirm the relationship between size and dissemination, but indicate a higher rate of small tumours presenting with widespread disease (Fig. 2). Early metastatic spread has long been recognized as an indicator for the poor prognosis of PDAC and, when detected, is a critical variable in the AJCC/TNM classification system on which treatment planning is based. Current clinical practice for PDAC recommends surgical resection of the primary tumour followed by adjuvant chemotherapy in patients with a resectable tumour. Based on the findings of the present study, early-stage tumours should also be considered high-risk lesions, with a high probability of systemic metastasis for which systemic treatment may be indicated. Such treatment would target micrometastases that go undetected by imaging. Accordingly, it was reported recently that patients with PDAC receiving neoadjuvant therapy before surgery showed clear improvement in median overall survival compared with patients treated by a ‘surgery-first’ approach29. The study has limitations given its retrospective design and the use of a national registry. The SEER historical stage A was used, which is different from the AJCC staging system and probably of less clinical relevance. However, this classification allowed uniform staging during the study period. Lack of central pathological review is a limitation of the SEER database. However, previous studies have found good agreement between the histological subtypes of cancer reported by SEER and those assigned by independent reviewers30. Another limitation of SEER data is related to registration of the use of chemotherapy. The completeness of the registration and the potential biases associated with reasons for receiving or not receiving chemotherapy have been addressed previously31. Missing values in this study were handled by a multiple imputation technique, which is probably the best method available today32. This approach reduces selection bias and improves generalizability, but necessitates caution when interpreting the results. In the future, the addition of specific tumour cell phenotypes may improve risk stratification for predicting metastatic disease and survival in patients with PDAC. Disclosure The authors declare no conflict of interest. Open in new tabDownload slide References 1 Siegel RL , Miller KD, Jemal A. Cancer statistics, 2015 . CA Cancer J Clin 2015 ; 65 : 5 – 29 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Ansari D , Tingstedt B, Andersson B, Holmquist F, Sturesson C, Williamsson C et al. Pancreatic cancer: yesterday, today and tomorrow . Future Oncol 2016 ; 12 : 1929 – 1946 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Hartwig W , Werner J, Jäger D, Debus J, Büchler MW. Improvement of surgical results for pancreatic cancer . Lancet Oncol 2013 ; 14 : e476 – e485 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Hruban RH , Maitra A, Goggins M. Update on pancreatic intraepithelial neoplasia . Int J Clin Exp Pathol 2008 ; 1 : 306 – 316 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 5 Yachida S , Jones S, Bozic I, Antal T, Leary R, Fu B et al. Distant metastasis occurs late during the genetic evolution of pancreatic cancer . Nature 2010 ; 467 : 1114 – 1117 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Yu J , Blackford AL, Dal Molin M, Wolfgang CL, Goggins M. Time to progression of pancreatic ductal adenocarcinoma from low-to-high tumour stages . Gut 2015 ; 64 : 1783 – 1789 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Furukawa H , Iwata R, Moriyama N. Growth rate of pancreatic adenocarcinoma: initial clinical experience . Pancreas 2001 ; 22 : 366 – 369 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Egawa S , Takeda K, Fukuyama S, Motoi F, Sunamura M, Matsuno S. Clinicopathological aspects of small pancreatic cancer . Pancreas 2004 ; 28 : 235 – 240 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Edge S , Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A. AJCC Cancer Staging Manual . Springer : New York , 2010 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 10 Sobin LH , Gospodarowicz MK, Wittekind C. The TNM Classification of Malignant Tumours . Wiley-Blackwell : Chichester , 2009 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 11 Fortner JG , Klimstra DS, Senie RT, Maclean BJ. Tumor size is the primary prognosticator for pancreatic cancer after regional pancreatectomy . Ann Surg 1996 ; 223 : 147 – 153 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Agarwal B , Correa AM, Ho L. Survival in pancreatic carcinoma based on tumor size . Pancreas 2008 ; 36 : e15 – e20 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Pongprasobchai S , Pannala R, Smyrk TC, Bamlet W, Pitchumoni S, Ougolkov A et al. Long-term survival and prognostic indicators in small (< or = 2 cm) pancreatic cancer . Pancreatology 2008 ; 8 : 587 – 592 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Franko J , Hugec V, Lopes TL, Goldman CD. Survival among pancreaticoduodenectomy patients treated for pancreatic head cancer < 1 or 2 cm . Ann Surg Oncol 2013 ; 20 : 357 – 361 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Rhim AD , Mirek ET, Aiello NM, Maitra A, Bailey JM, McAllister F et al. EMT and dissemination precede pancreatic tumor formation . Cell 2012 ; 148 : 349 – 361 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Ishikawa O , Ohigashi H, Imaoka S, Nakaizumi A, Uehara H, Kitamura T et al. Minute carcinoma of the pancreas measuring 1 cm or less in diameter – collective review of Japanese case reports . Hepatogastroenterology 1999 ; 46 : 8 – 15 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 17 National Cancer Institute Surveillance, Epdemiology, and End Results Program . SEER*Stat Database: Incidence – SEER 18 Regs Custom Data (with chemotherapy recode), Nov 2015 Sub (1973–2013 varying) . https://seer.cancer.gov/seerstat/ [accessed 25 July 2016 ]. 18 Vandenbroucke JP , von Elm E, Altman DG, Gøtzsche PC, Mulrow CD, Pocock SJ et al. ; STROBE Initiative. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration . Epidemiology 2007 ; 18 : 805 – 835 . Google Scholar Crossref Search ADS PubMed WorldCat 19 White IR , Royston P, Wood AM. Multiple imputation using chained equations: issues and guidance for practice . Stat Med 2011 ; 30 : 377 – 399 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Geer RJ , Brennan MF. Prognostic indicators for survival after resection of pancreatic adenocarcinoma . Am J Surg 1993 ; 165 : 68 – 72 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Willett CG , Czito BG, Bendell JC, Ryan DP. Locally advanced pancreatic cancer . J Clin Oncol 2005 ; 23 : 4538 – 4544 . Google Scholar Crossref Search ADS PubMed WorldCat 22 de Jong MC , Li F, Cameron JL, Wolfgang CL, Edil BH, Herman JM et al. Re-evaluating the impact of tumor size on survival following pancreaticoduodenectomy for pancreatic adenocarcinoma . J Surg Oncol 2011 ; 103 : 656 – 662 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Hishinuma S , Ogata Y, Tomikawa M, Ozawa I, Hirabayashi K, Igarashi S. Patterns of recurrence after curative resection of pancreatic cancer, based on autopsy findings . J Gastrointest Surg 2006 ; 10 : 511 – 518 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Iacobuzio-Donahue CA , Fu B, Yachida S, Luo M, Abe H, Henderson CM et al. DPC4 gene status of the primary carcinoma correlates with patterns of failure in patients with pancreatic cancer . J Clin Oncol 2009 ; 27 : 1806 – 1813 . Google Scholar Crossref Search ADS PubMed WorldCat 25 Bailey P , Chang DK, Nones K, Johns AL, Patch AM, Gingras MC et al. Genomic analyses identify molecular subtypes of pancreatic cancer . Nature 2016 ; 531 : 47 – 52 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Hruban RH , Goggins M, Parsons J, Kern SE. Progression model for pancreatic cancer . Clin Cancer Res 2000 ; 6 : 2969 – 2972 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 27 Real FX , Cibrián-Uhalte E, Martinelli P. Pancreatic cancer development and progression: remodeling the model . Gastroenterology 2008 ; 135 : 724 – 728 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Haeno H , Gonen M, Davis MB, Herman JM, Iacobuzio-Donahue CA, Michor F. Computational modeling of pancreatic cancer reveals kinetics of metastasis suggesting optimum treatment strategies . Cell 2012 ; 148 : 362 – 375 . Google Scholar Crossref Search ADS PubMed WorldCat 29 Christians KK , Heimler JW, George B, Ritch PS, Erickson BA, Johnston F et al. Survival of patients with resectable pancreatic cancer who received neoadjuvant therapy . Surgery 2016 ; 159 : 893 – 900 . Google Scholar Crossref Search ADS PubMed WorldCat 30 Field RW , Smith BJ, Platz CE, Robinson RA, Neuberger JS, Brus CP et al. Lung cancer histologic type in the Surveillance, Epidemiology, and End Results registry versus independent review . J Natl Cancer Inst 2004 ; 96 : 1105 – 1107 . Google Scholar Crossref Search ADS PubMed WorldCat 31 Noone AM , Lund JL, Mariotto A, Cronin K, McNeel T, Deapen D et al. Comparison of SEER treatment data with Medicare claims . Med Care 2016 ; 54 : e55 – e64 . Google Scholar Crossref Search ADS PubMed WorldCat 32 Sterne JA , White IR, Carlin JB, Spratt M, Royston P, Kenward MG et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls . BMJ 2009 ; 338 : b2393 . Google Scholar Crossref Search ADS PubMed WorldCat © 2017 BJS Society Ltd Published by John Wiley & Sons Ltd This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © 2017 BJS Society Ltd Published by John Wiley & Sons Ltd
TI - Relationship between tumour size and outcome in pancreatic ductal adenocarcinoma
JO - British Journal of Surgery
DO - 10.1002/bjs.10471
DA - 2017-03-14
UR - https://www.deepdyve.com/lp/oxford-university-press/relationship-between-tumour-size-and-outcome-in-pancreatic-ductal-4dYz3LE0TR
SP - 600
EP - 607
VL - 104
IS - 5
DP - DeepDyve
ER -