Impact of anatomic site of biliary tract tumour origin and conditional probability of survival (CS): results from 15 prospective advanced first-line clinical trials

Background: Inclusion of all patients (pts) with advanced biliary tract cancer (aBTC), irrespective of anatomic location, with assessment of overall survival (OS) in prospective trials, is debated. Additionally, outcome is typically described as estimated OS, but CS offers more relevant information once pts have survived for some time; this study assessed the impact of anatomic site of BTC origin and CS. Methods: Pts enrolled into 15 prospective first-line aBTC clinical trials were included. OS was analysed using Cox proportional hazard regression; CS and 95% confidence intervals (CIs) were calculated. Results: Overall, 1333 pts were included (Jan 97-Dec 13) with a median (med) age of 63 yrs (range 23-85); 46% male; 84% ECOG PS 0/1; 25% locally advanced (LA) stage, 72% metastatic (met), and 3% not reported (NR). Anatomic site of origin was gallbladder (GBC): 385 (29%), cholangiocarcinoma not specified (CCA-NS): 363 (27%), extrahepatic (EHC): 247 (19%), intrahepatic (IHC): 209 (16%), ampulla: 53 (4%) and 76 (6%) NR. Treatment was mono-chemotherapy: 310 (23%), cis/gem combination: 482 (36%), other combination: 520 (39%) and NR: 21 (2%). Med OS: 10.2 mths (95% CI 9.6-10.9). All sites, adjusted for treatment, had decreased risk of death vs GBC: EHC (p<.001), IHC (p<.002), CCA-NS (p<.003), ampulla (p=.003). This reduced risk vs GBC was maintained in those receiving cis/gem combination therapy in EHC (p<.001) and IHC (p<.001), but not in CCA-NS (p=.82) or ampulla (p=.96). Probabilities of surviving an additional yr given survival to 1 (n=552), 2 (n=170), 3 (n=53), and 4 (n=23) yrs post trial registration were 37%, 45%, 61%, and 63% respectively. For pts who survived 1 yr; those receiving combination therapy vs mono (p=.008), and those with IHC and CCA-NS vs GBC had better CS (both p<.05). Met stage vs LA was associated with shorter CS (p=.002) and ECOG PS and gender had no effect on CS (p>.05, p=.08 respectively). Conclusions: Pts with GBC have worse OS compared to other anatomic BTC sites. Inclusion of other BTC subtypes, at least, in prospective aBTC clinical trials is justified. Conditional probabilities allow adjusted prognosis prediction for survivors with aBTC.