The impact of viral etiology in yttrium-90–treated hepatocellular carcinoma

608

Background: Hepatocellular carcinoma (HCC) is the most common type of liver cancer and is derived from a background of chronic inflammation from both viral and environmental factors. Recent data indicates that immune therapy efficacy is improved in virally mediated HCC. We sought to investigate what impact HCC etiology has on outcomes in patients treated with yttrium-90 (Y90). Methods: We conducted a retrospective review of patients with HCC treated at our institution with Y90 radiotherapy from 2005-2021. Clinical and treatment characteristics were recorded, including: age, stage, HCC etiology, performance status, Child Pugh class, liver mass/volume treated, dose, prior therapy, imaging characteristics and basic labs. Univariable (UVA) and multivariable analyses (MVA) were conducted to identify prognostic factors for local control (LC) and overall survival (OS) following Y90. Results: 67 patients and 196 distinct tumors treated with Y90 were identified. Median prescription dose was 110 Gy (range 44-157). Patients were stage I-IVB with the most frequent stages being II, IIIb and IIIa (34%, 21% and 19% patients, respectively). 49 patients (73%) were Child-Pugh A and 18 (27%) were Child-Pugh B. 5 patients only had hepatitis B (7%), 26 patients only had hepatitis C (39%), and 6 patients had both hepatitis B and C (9%). Median OS among all patients was 7.6 (95% CI 5.6-11.5) months following Y90. The LC rate of all treated lesions was 42% and 24% at 6 and 12 months, respectively. On UVA analysis, improved LC was associated with younger age ( p=0.019), smaller liver mass ( p=0.039), and Hepatitis B etiology ( p=0.018). On Cox regression survival analysis, worsened OS was associated with lower ECOG status ( p=0.024), advanced stage ( p=0.023), lower albumin ( p=0.001), higher AFP ( p<0.001), and portal vein invasion/thrombosis ( p=0.006, p=0.010). On MVA, lower albumin, higher AFP, advance stage and prior sorafenib were significantly related to worsened OS. If hepatitis B or C was forced into the MVA, the hazard ratio was 0.6 (95% CI 0.3-1.04) with p=0.068. Conclusions: Our results indicate that viral etiology impacts LC in our patients. Larger studies are necessary to confirm these results and to determine whether viral etiology impacts OS.

Germline screening rates and patterns for patients with pancreatic cancer at an academic medical center

10590

Background: Current National Comprehensive Cancer Network guidelines recommend germline genetic testing for all pancreatic cancer patients irrespective of family history. Germline testing provides insight on inherited pathogenic variants that may influence care. Approximately 10% of pancreatic adenocarcinoma (PDAC) patients have pathogenic mutations which may have treatment implications and warrant the introduction of targeted therapy approaches. Patients with germline BRCA1/2 or PALB2 mutations have increased sensitivity to platinum chemotherapy and PARP-inhibitor therapies. Germline testing results may have important implications for patients’ family members for earlier targeted screening. A better understanding of the current state of testing is needed to develop systems to improve screening rates. We conducted a retrospective review of clinical practice patterns at an academic cancer center to assess the current uptake. Methods: Patients with pancreatic adenocarcinoma seen at the University of Virginia Health System within the 2021 calendar year were identified. Retrospective review of genetic counseling referral and germline genomic screening for individual patients was performed. Results: 210 patients with pancreatic adenocarcinoma were identified. 39 (19%) PDAC patients had a referral to genetic counseling placed in the electronic medical record and 44 (21%) completed germline screening. Of the patients referred to genetics, 17/39 patients (44%) met with a genetic counselor which led to germline screening, 3/39 (8%) patients were referred and saw genetics after receiving germline testing results. Among patients who completed germline screening, 27/44 (61%) had testing initially ordered by their oncologist with referral to genetic counseling based on testing results 3/27 (11%). Conclusions: Despite guideline recommendations, germline testing rates are low among this PDAC population. Genetic counselors are essential members of a multidisciplinary team and guide patient discussions and decision making with regards to germline testing. Typical practice has involved referral to meet with a genetics counselor prior to testing; however many patients elect not to schedule a visit and consequently do not obtain germline screening. Barriers may include costs associated with genetic counseling/testing, time constraints, and patient understanding of the relevance of testing for their cancer care. We observed that offering germline testing to PDAC patients with referral to genetic counseling based on results and patient preference is a viable practice pattern. Upfront clinician driven germline testing may offer an opportunity to improve access to germline screening. Prospective clinical trials are needed to increase rates of germline testing and genetic counseling for PDAC patients.

A phase 1b multitumor cohort study of cabozantinib plus atezolizumab in advanced solid tumors (COSMIC-021): Results of the colorectal cancer cohort

121

Background: Cabozantinib, a multiple receptor tyrosine kinase inhibitor, promotes an immune-permissive environment which may enhance the activity of immune checkpoint inhibitors. COSMIC-021 (NCT03170960) is evaluating the combination of cabozantinib with atezolizumab, an anti-PD-L1 inhibitor, in patients with advanced solid tumors. Outcomes in patients (pts) with metastatic colorectal cancer (mCRC) previously treated with fluoropyrimidine-containing therapy are presented. Methods: Pts with mCRC and an ECOG PS of 0–1 who progressed during or following systemic chemotherapy including fluoropyrimidine plus oxaliplatin or irinotecan were eligible. Up to 2 prior lines of anti-cancer therapy including EGFR-targeted therapy were allowed. Microsatellite instability high (MSI-H) and/or mismatch repair (MMR)-deficient pts were excluded. Pts received cabozantinib 40 mg PO QD plus atezolizumab 1200 mg IV Q3W. The primary endpoint was objective response rate (ORR) per RECIST 1.1 by investigator. Other endpoints included safety, duration of response (DOR), progression-free survival (PFS), and overall survival (OS). CT/MRI scans were performed Q6W for the first year and Q12W thereafter. Results: 31 pts received cabozantinib plus atezolizumab (median age, 60 y [range 31, 79]; male, 58%; ECOG PS 1, 61%; 2 prior lines of therapy, 71%; prior EGFR inhibitor, 16%; ≥3 tumor sites, 52%; tumors in left colorectum, 71%). Median follow-up was 28.1 mo (range, 24.2, 31.3) as of July 21, 2021. Cabozantinib plus atezolizumab demonstrated clinical activity in pts with mCRC (Table). Patients with wild-type RAS (n = 12) had numerically longer PFS and OS and higher ORR vs those with mutations (n = 19) (Table). Treatment-related adverse events (TRAEs) of any grade occurred in 28 (90%); the most common were diarrhea (52%), fatigue (42%), and nausea (35%). Grade 3-4 TRAEs occurred in 16 (52%); the most common were hypertension (10%), fatigue (6%), and lipase increased (6%); no Grade 5 events were reported. Conclusions: Cabozantinib plus atezolizumab demonstrated encouraging clinical activity with manageable toxicity in pts with previously treated advanced non-MSI-H/MMR-proficient CRC. Clinical trial information: NCT03170960. [Table: see text]

Institutional experience with total neoadjuvant therapy for low-lying rectal tumors

123

Background: Surgery has traditionally been the primary treatment for rectal cancer with consideration of neoadjuvant and adjuvant therapies based on stage. Recent studies have shown that Total Neoadjuvant Therapy (TNT) may provide excellent outcomes, with a possibility of non-operative management in patients with rectal cancer. This study is an initial report of our institution’s watchful waiting strategy. Methods: Patients with Stage II-III, low rectal adenocarcinomas treated from 2015-2018 with TNT were retrospectively reviewed. Patients were surgical candidates with no planned surgery. All patients received definitive radiation (median dose 54 Gy, range 50-56 Gy, at 1.8-2 Gy/fx) with concurrent Xeloda, with additional chemotherapy delivered either prior to or following chemoradiation. Kaplan-Meier (KM) method was utilized to estimate the 2 year permanent ostomy and surgery free survivals. Univariate and multivariate analysis using binary logistic regression were performed to assess the odds ratio (OR) of the need for surgery, with 95% confidence interval (CI). Results: 28 patients were treated with TNT with a median age of 59 years (range 32-79 yrs) and median follow up of 24 months (range 6-51 mon). Reasons for TNT included: clinical trial (50%, n = 14), patient desire to avoid surgery (43%, n = 12), and history of LDR prostate brachytherapy (7%, n = 2). The majority of patients had Stage III disease (68%, n = 19). Median tumor distance from the anal verge was 3 cm (range 0-7 cm). 75% (n = 21) of patients were initially managed without surgery, with 5 patients requiring LAR and 2 requiring APR for residual disease. 4 patients initially treated without surgery experienced local recurrence (LR) requiring APR, with a median time to LR of 20 months (range 12-27 mon). The KM estimated 2 year permanent ostomy and surgery free survivals were 69% and 56% respectively. Distant metastases occurred in two patients at 2 and 7 months post RT. An incomplete response on post-treatment MRI predicted eventual receipt of surgery (p = 0.012, OR = 19.8, 95% CI 1.9-202, compared to complete responders). Conclusions: These results support the growing evidence that TNT may provide a non-surgical option for select patients with low lying rectal cancer.

Clinical and immune responses using anti-CD3 x anti-EGFR bispecific antibody armed T cells (BATs) for locally advanced or metastatic pancreatic cancer

4135

Background: Conventional chemotherapy (chemo) for locally advanced pancreatic cancer (LAPC) and metastatic pancreatic cancer (MPC) has dismal responses and poor survival rates. Arming activated T cells (ATC) with anti-CD3 x anti-EGFR bispecific antibody (BATs) makes every ATC into an EGFR-specific cytotoxic T cell that secretes cytokines, proliferates, and kills tumor. Methods: We report on 5 phase I (P1) and 15 phase II (P2) patients. In our phase I study, BATs were used to treat LAPC or MPC patients at Karmanos Cancer Institute (NCT0140874) in a dose escalation involving 3 weekly infusions of 1, 2, and 4 x 1010 BATs/infusion, followed by a booster infusion at 3 months (mos) for a total of up to 8 x 1010 BATs. No dose limiting toxicities were observed in the outpatient infusions. Fifteen patients treated on a phase II (NCT02620865) at KCI and (NCT03269526) at University of Virginia received biweekly infusions of 1010 BATs/infusion over 4 weeks for a total of 8 x 1010 EGFR BATs. Results: Four patients had stable disease (SD) for 6.1, 6.5, 5.3, and 36 mos. Two patients had complete responses (CR) when chemo was restarted after BATs. The median overall survival (OS) for 17 evaluable patients (3 of 4 infusions in the P1 and all 8 infusions in the P2) was 31 mos, and the median OS for all 20 patients (3 in the P2 who did not complete 8 infusions) is 14.5 mos (95% CI, 7.5-45.2 mos). Patient IT20104 had an apparent “pseudoprogression” after 3 BATs infusions, but achieved a CR after restarting capcitabine and is alive off therapy at 54 mos (24 mos after stopping capecitabine). Immune evaluations on the P1 patients show specific cytotoxicity to MiaPaCa-2 by peripheral blood mononuclear cells (PBMC) increased from 21% to 31% 2 weeks after the 3rd infusion, and IFN-γ EliSpots increased from < 20 to 1000 IFN-γ EliSpots/106 PBMC (p < 0.03). Patient IT 20121 (SD for 36 mos) increased IFN-γ EliSpots from 250 to 3200/106 PBMC after 8 infusions. Innate cytotoxicity responses in the P1 patients increased significantly after infusions (p < 0.04). Levels of IP-10 increased significantly (p < 0.04), and levels of IL-8 decreased but not significantly (p < 0.07). Conclusions: Infusions of BATs are safe and induce endogenous adaptive anti-tumor responses. Targeting PC with BATs may stabilize disease, leading to improved OS, as well as evidence that BATs infusions can induce anti-tumor activity and immunosensitize tumors to subsequent chemo. Clinical trial information: NCT014084,NCT03269526,NCT02620865.

Mismatch repair deficiency in cholangiocarcinoma

269

Background: Cholangiocarcinoma (CC) is a fatal malignancy with an unmet treatment need. With the approval of immunotherapy for solid tumors with mismatch repair (MMR) deficiency, there is a renewed interest in MMR testing. Little is known about the incidence of MMR deficiency in CC or its correlation to survival, immune cell infiltration, PD-L1 and other proteins expressed in CC such as mesothelin. Methods: CC tumors were identified from patients treated at the University of Virginia from 2000-2014. Tissue microarrays (TMAs) were constructed of 3-4 cores from each tumor and were stained by immunohistochemistry for MMR genes (MLH1, PMS2, MSH2, MSH6), mesothelin, PD-L1 and immune cells. TMAs were scanned using the Leica SCN400 and analyzed using the Digital Image Hub software. Stain intensity thresholds for defining positive cells were determined by two users and recorded as an average of all cores from each tumor. Mesothelin and PD-L1 expression were measured as a percentage of positive tumor cells. Correlation with overall survival was assessed using log-rank tests and classification and regression trees, with p values < 0.05 considered significant. Results: Ninety-one tumors were analyzed: 24 intrahepatic, 33 hilar, and 34 distal. MMR deficiency was found in 20 tumors (22%). None of the MMR deficient tumors co-expressed PD-L1 (>1%), which was found in 15% of the remaining tumors. T cell infiltration (CD4, CD8 and FoxP3) did not differ between MMR deficient or proficient tumors. Patients with MMR deficiency had a trend towards worse survival compared to those with proficiency (median OS: 19.2 vs. 28.1 months, p = 0.07). MMR deficient tumors showed a lower mesothelin expression compared to MMR proficient tumors, median 8 vs. 129 positive cells per TMA (p = 0.08). Patients with MMR deficiency and low mesothelin expression had a worse outcome compared to patients with MMR proficiency and high mesothelin expression (median OS: 14.5 vs. 30.0 months, p = 0.05). Conclusions: Given the high rate of MMR deficiency, all CC tumors should be tested and may benefit from anti-PD-1 therapy. The poor prognosis of MMR deficient CC may be independent of T-cell infiltration and additional studies are needed to better characterize the genetic and molecular landscape of this subset of tumors.

Correlation of mesothelin expression and CD8 tumor infiltrating lymphocytes with prognosis in cholangiocarcinoma

e15650

Background: Cholangiocarcinoma (CC) is a rapidly progressing malignancy with an unmet treatment need. Little is known about the CC tumor immune microenvironment or about relevant antigenic targets. We hypothesized that lack of T cell infiltration or PD-L1 expression may identify patients at high risk of death, and that mesothelin may be a relevant antigenic target. Methods: A retrospective analysis was conducted of CC tumors at the University of Virginia from 2000-2014. TMAs were constructed of 3-4 cores from each tumor and were stained by IHC for CD4 and CD8 tumor infiltrating lymphocytes (TILs), mesothelin and PD-L1. TMAs were scanned using the Leica SCN400 and analyzed using the Digital Image Hub software. Stain intensity thresholds for defining positive cells were determined by two users and recorded as an average of all cores from each tumor. Mesothelin and PD-L1 expression were measured as a percentage of positive tumor cells. TILs and protein expression were analyzed for association with overall survival, grouped as high or low expression based either on the median or the 33rdpercentile. Correlation with overall survival was assessed using a log rank test and a classification and regression tree with p-values < 0.05 being considered statistically significant. Results: Ninety-nine tumors were available for analysis: 26 intrahepatic, 37 hilar, and 36 distal. PD-L1 and mesothelin expression > 1% of tumor cells were found in 16% and 92% of tumors, respectively. CD4 and CD8 TILs were found in nearly all tumors (98% and 96%), with the majority showing intraepithelial CD4 and CD8 infiltration (73% and 68%). There were no significant associations between survival and PD-L1, mesothelin, or CD4 and CD8 infiltration. However when considered together, the group with low mesothelin/low CD8 (each below 33rdpercentile) had worse survival (9.1 months) compared to high mesothelin/high CD8 (25 months), high mesothelin/low CD8 (30.1 months) and low mesothelin/high CD8 (26.1 months), p = 0.015. Conclusions: CC tumors that lack CD8 infiltration and mesothelin expression have a poor prognosis. Mesothelin represents an attractive target in cholangiocarcinoma, opening the door for future immunotherapy for CC.