Immunological aspects on prognosis in resectable cholangiocellular cancer

Abstract

Background: Biliary tract cholangiocellular cancers, cholangiocarcinoma (CCA) and gallbladder cancer (GBC), are malignancies with poor prognosis. Even after resection surgery with curative intent, a majority of patients suffer recurrence, and median overall survival (OS) remains limited to approximately three years. To improve long-term survival, a better understanding of prognostic factors is needed to individualize treatment.

Aims: Paper I: To evaluate the prognostic value of two preoperative inflammation-based prognostic scores, the Glasgow prognostic score (GPS) and the Modified Glasgow prognostic score (mGPS). Paper II: To evaluate prognostic factors and outcomes after hepatobiliary resection for perihilar CCA (pCCA) in patients with underlying primary sclerosing cholangitis (PSC), a chronic hepatobiliary inflammatory condition. Paper III: To identify specific immunologic prognostic markers and to further characterize the immune response in resectable biliary tract cancer. Paper IV: To systematically review the prognostic influence of multiple hepatic lesions in patients undergoing resection for intrahepatic CCA (iCCA), with stratification according to distribution and number of lesions.

Methods: Paper I was a retrospective single-centre study, including patients undergoing surgery for iCCA, pCCA or GBC (Karolinska University Hospital 2009-2017). The primary outcome was OS, secondary outcome complications. Survival was analysed by the Kaplan- Meier method and Cox regression. Paper II was a retrospective multi-centre cohort study including patients undergoing resection for pCCA at 21 centres (Europe, United States 2000-2020). The primary outcome variable was OS, secondary outcomes disease-free survival and postoperative complications. Survival was analysed by Kaplan-Meier method and Cox regression. Paper III was a retrospective single-centre cohort study, including patients undergoing surgery for suspected biliary tract cancer at Karolinska University Hospital (2009-2017). The primary outcome variable was OS. Plasma expression of immune-related proteins was analysed in prospectively collected biobank samples by Proximity Extension Assay. Survival associations were analysed by Cox regression. Tissue expression of identified markers and receptors/ligands was analysed in independent public cohorts. Paper IV was a systematic review and meta-analysis (Medline [Ovid] and Embase, 2010- 2021). Original articles with data on OS stratified for tumour distribution (satellite lesions/other multiple lesions) and/or tumour number were included. The study was preregistered in a public prospective register of systematic reviews and PRISMA 2020 reporting guidelines were followed.

Results: In paper I, the GPS and the mGPS were independent prognostic factors for overall survival after resection for biliary tract cancer (GPS≥1 hazard ratio [HR] 2.35, 95% confidence interval [CI] 1.41-3.93, mGPS≥1 HR 1.68, 95% CI 1.05-2.68). The GPS, but not the mGPS, identified an intermediate risk group. In paper II, median OS was 33 months (95% CI 10-54 months) for patients with PSCassociated pCCA and 29 months (95% CI 26-32 months) for patients without underlying PSC. Patients with PSC-associated pCCA had a lower rate of well-differentiated tumours (3% vs. 16%, p=0.043), a higher rate of postoperative complications (71% vs. 44%, p=0.003) and similar 90-day mortality (12% vs 13%, p=1.00). In paper III, three proteins in preoperative plasma were independently associated with OS: TRAIL/TNFSF10 (HR 0.30, 95% CI 0.16-0.56), TIE2/TEK (HR 2.78, 95% CI 1.20-6.48) and CSF1/M-CSF (HR 4.02, 95% CI 1.40-11.59). TRAIL/TNFSF10 was a positive prognostic factor in iCCA and pCCA. CSF1/M-CSF was a negative prognostic factor in iCCA and GBC. TIE2/TEK was a negative prognostic factor in GBC. In CCA tissue analysis, TRAIL-R1/TNFSFR10A receptor expression was higher in tumour and TRAIL/TNFSF10 was expressed by intratumoral lymphocytes, NK-cells and monocytes. CSF1/M-CSF was expressed by tumour-infiltrating CD8+ T-cells. In paper IV, OS was decreased for iCCA patients with satellite lesions (HR 1.89, 95% CI 1.67-2.13) and multiple lesions other than satellites (HR 2.41, 95% CI 1.72-3.37). Data stratified for tumour number was limited, but indicated increased risk per additional lesion.

Conclusions: Preoperative systemic inflammatory markers were independent prognostic factors for OS after resection for BTC. Median OS after resection for pCCA was similar for patients with and without underlying PSC. Patients with PSC had a lower rate of well-differentiated tumours and a higher rate of complications. Three specific immunological protein markers in preoperative plasma were associated with OS in BTC, with disease-specific differences on subgroup analysis. iCCA-specific markers TRAIL/TNFSF10 and CSF1/M-CSF were expressed in tumour-infiltrating immune-cells. Satellite lesions, as well as multiple lesions other than satellites were negative prognostic factors in iCCA. The number of lesions was suggested to be a prognostic factor within the multiple lesion group.