Large Language Models for Automated Synoptic Reports and Resectability Categorization in Pancreatic Cancer.

Background Structured radiology reports for pancreatic ductal adenocarcinoma (PDAC) improve surgical decision-making over free-text reports, but radiologist adoption is variable. Resectability criteria are applied inconsistently. Purpose To evaluate the performance of large language models (LLMs) in automatically creating PDAC synoptic reports from original reports and to explore performance in categorizing tumor resectability. Materials and Methods In this institutional review board-approved retrospective study, 180 consecutive PDAC staging CT reports on patients referred to the authors' European Society for Medical Oncology-designated cancer center from January to December 2018 were included. Reports were reviewed by two radiologists to establish the reference standard for 14 key findings and National Comprehensive Cancer Network (NCCN) resectability category. GPT-3.5 and GPT-4 (accessed September 18-29, 2023) were prompted to create synoptic reports from original reports with the same 14 features, and their performance was evaluated (recall, precision, F1 score). To categorize resectability, three prompting strategies (default knowledge, in-context knowledge, chain-of-thought) were used for both LLMs. Hepatopancreaticobiliary surgeons reviewed original and artificial intelligence (AI)-generated reports to determine resectability, with accuracy and review time compared. The McNemar test, t test, Wilcoxon signed-rank test, and mixed effects logistic regression models were used where appropriate. Results GPT-4 outperformed GPT-3.5 in the creation of synoptic reports (F1 score: 0.997 vs 0.967, respectively). Compared with GPT-3.5, GPT-4 achieved equal or higher F1 scores for all 14 extracted features. GPT-4 had higher precision than GPT-3.5 for extracting superior mesenteric artery involvement (100% vs 88.8%, respectively). For categorizing resectability, GPT-4 outperformed GPT-3.5 for each prompting strategy. For GPT-4, chain-of-thought prompting was most accurate, outperforming in-context knowledge prompting (92% vs 83%, respectively; P = .002), which outperformed the default knowledge strategy (83% vs 67%, P < .001). Surgeons were more accurate in categorizing resectability using AI-generated reports than original reports (83% vs 76%, respectively; P = .03), while spending less time on each report (58%; 95% CI: 0.53, 0.62). Conclusion GPT-4 created near-perfect PDAC synoptic reports from original reports. GPT-4 with chain-of-thought achieved high accuracy in categorizing resectability. Surgeons were more accurate and efficient using AI-generated reports. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Chang in this issue.

Disease-free survival after pancreatectomy for pancreatic neuroendocrine tumours: a 17-year single-center experience of 223 patients.

BACKGROUND: Metastasis or recurrence of pancreatic neuroendocrine tumours (pNETs) after pancreatectomy presents an important source of post-surgical morbidity. Accordingly we aim to define disease-free survival (DFS) in this population. METHODS: Patients who underwent pancreatectomy for pNETs between Jan 2005 and Jan 2022, were included. Clinicopathologic and survival data were collected and the primary endpoint was DFS. Kaplan-Meier survival analysis and Cox proportional hazards regression modelling were performed. RESULTS: Of the 223 patients, 144 (65%) distal/subtotal/partial pancreatectomies, 71 (32%) pancreaticoduodectomies, 6 (3%) total pancreatectomies, and 2 (1%) enucleations were performed. Forty-five (20%) patients had disease recurrence or metastasis after index pancreatectomy during the 17 years of the study. Non-functional pNETs (n=162, 73%) were more common than hormonally functional sub-types. The 2-year and 5-year DFS were 82% and 76%, respectively. Kaplan-Meier analysis demonstrated node positivity, size ≥4cm, lymphovascular invasion, perineural invasion, Ki-67 ≥20% and non-functional pNETs to be significantly associated with lowered DFS (p<0.05). Univariate Cox analysis identified the following predictors to be significantly associated with poorer DFS: larger tumour size (HR: 1.16 [95% CI: 1.04-1.28]), Ki-67 index ≥20% (HR: 4.93 [95% CI: 2.00-11.44]), perineural invasion (HR: 3.23 [95% CI: 1.40-7.89]), open surgery (HR: 3.34 [95% CI: 1.03-1.33]), node-positive disease (HR: 5.27 [95% CI: 2.28-13.26]) and increased BMI (HR: 1.10 [95% CI: 1.03-1.17]) (p<0.05). CONCLUSIONS: One in every five resected patients developed recurrence or metastasis after pancreatectomy. Prognostic predictors of DFS in pNETs could help optimize treatment and enhance follow-up protocols to improve quality and reduce morbidity.

In vivo CRISPR screens reveal SCAF1 and USP15 as drivers of pancreatic cancer.

Functionally characterizing the genetic alterations that drive pancreatic cancer is a prerequisite for precision medicine. Here, we perform somatic CRISPR/Cas9 mutagenesis screens to assess the transforming potential of 125 recurrently mutated pancreatic cancer genes, which revealed USP15 and SCAF1 as pancreatic tumor suppressors. Mechanistically, we find that USP15 functions in a haploinsufficient manner and that loss of USP15 or SCAF1 leads to reduced inflammatory TNFα, TGF-ß and IL6 responses and increased sensitivity to PARP inhibition and Gemcitabine. Furthermore, we find that loss of SCAF1 leads to the formation of a truncated, inactive USP15 isoform at the expense of full-length USP15, functionally coupling SCAF1 and USP15. Notably, USP15 and SCAF1 alterations are observed in 31% of pancreatic cancer patients. Our results highlight the utility of in vivo CRISPR screens to integrate human cancer genomics and mouse modeling for the discovery of cancer driver genes with potential prognostic and therapeutic implications.

Genome-wide analysis to assess if heavy alcohol consumption modifies the association between SNPs and pancreatic cancer risk.

BACKGROUND: Pancreatic cancer is a leading cause of cancer-related death globally. Risk factors for pancreatic cancer include common genetic variants and potentially heavy alcohol consumption. We assessed if genetic variants modify the association between heavy alcohol consumption and pancreatic cancer risk. METHODS: We conducted a genome-wide interaction analysis of single nucleotide polymorphisms (SNP) by heavy alcohol consumption (more than 3 drinks per day) for pancreatic cancer in European ancestry populations from genome-wide association studies (GWAS). Our analysis included 3,707 cases and 4,167 controls from case-control studies and 1,098 cases and 1,162 controls from cohort studies. Fixed effect meta-analyses were conducted. RESULTS: A potential novel region of association on 10p11.22, lead SNP rs7898449 (Pinteraction = 5.1 x 10-8 in the meta-analysis, Pinteraction = 2.1x10-9 in the case-control studies, Pinteraction = 0.91 cohort studies) was identified. A SNP correlated with this lead SNP is an eQTL for the NRP1 gene. Of the 17 genomic regions with genome-wide significant evidence of association with pancreatic cancer in prior studies, we observed suggestive evidence that heavy alcohol consumption modified the association for one SNP near LINC00673, rs11655237 on 17q25.1 (Pinteraction = 0.004). CONCLUSIONS: We identified a novel genomic region that may be associated with pancreatic cancer risk in conjunction with heavy alcohol consumption located near an eQTL for the NRP1, a protein that plays an important role in the development and progression of pancreatic cancer Impact: This work can provide insight into the etiology of pancreatic cancer particularly in heavy drinkers.

Use of Deep Learning to Evaluate Tumor Microenvironmental Features for Prediction of Colon Cancer Recurrence.

Deep learning may detect biologically important signals embedded in tumor morphologic features that confer distinct prognoses. Tumor morphologic features were quantified to enhance patient risk stratification within DNA mismatch repair (MMR) groups using deep learning. Using a quantitative segmentation algorithm (QuantCRC) that identifies 15 distinct morphologic features, we analyzed 402 resected stage III colon carcinomas [191 deficient (d)-MMR; 189 proficient (p)-MMR] from participants in a phase III trial of FOLFOX-based adjuvant chemotherapy. Results were validated in an independent cohort (176 d-MMR; 1,094 p-MMR). Association of morphologic features with clinicopathologic variables, MMR, KRAS, BRAFV600E, and time-to-recurrence (TTR) was determined. Multivariable Cox proportional hazards models were developed to predict TTR. Tumor morphologic features differed significantly by MMR status. Cancers with p-MMR had more immature desmoplastic stroma. Tumors with d-MMR had increased inflammatory stroma, epithelial tumor-infiltrating lymphocytes (TIL), high-grade histology, mucin, and signet ring cells. Stromal subtype did not differ by BRAFV600E or KRAS status. In p-MMR tumors, multivariable analysis identified tumor-stroma ratio (TSR) as the strongest feature associated with TTR [HRadj 2.02; 95% confidence interval (CI), 1.14-3.57; P = 0.018; 3-year recurrence: 40.2% vs. 20.4%; Q1 vs. Q2-4]. Among d-MMR tumors, extent of inflammatory stroma (continuous HRadj 0.98; 95% CI, 0.96-0.99; P = 0.028; 3-year recurrence: 13.3% vs. 33.4%, Q4 vs. Q1) and N stage were the most robust prognostically. Association of TSR with TTR was independently validated. In conclusion, QuantCRC can quantify morphologic differences within MMR groups in routine tumor sections to determine their relative contributions to patient prognosis, and may elucidate relevant pathophysiologic mechanisms driving prognosis. SIGNIFICANCE: A deep learning algorithm can quantify tumor morphologic features that may reflect underlying mechanisms driving prognosis within MMR groups. TSR was the most robust morphologic feature associated with TTR in p-MMR colon cancers. Extent of inflammatory stroma and N stage were the strongest prognostic features in d-MMR tumors. TIL density was not independently prognostic in either MMR group.

Impact of an Inter-Professional Clinic on Pancreatic Cancer Outcomes: A Retrospective Cohort Study.

Background: Pancreatic ductal adenocarcinoma (PDAC) presents significant challenges in diagnosis, staging, and appropriate treatment. Furthermore, patients with PDAC often experience complex symptomatology and psychosocial implications that require multi-disciplinary and inter-professional supportive care management from health professionals. Despite these hurdles, the implementation of inter-professional clinic approaches showed promise in enhancing clinical outcomes. To assess the effectiveness of such an approach, we examined the impact of the Wallace McCain Centre for Pancreatic Cancer (WMCPC), an inter-professional clinic for patients with PDAC at the Princess Margaret Cancer Centre (PM). Methods: This retrospective cohort study included all patients diagnosed with PDAC who were seen at the PM before (July 2012-June 2014) and after (July 2014-June 2016) the establishment of the WMCPC. Standard therapies such as surgery, chemotherapy, and radiation therapy remained consistent across both time periods. The cohorts were compared in terms of survival rates, disease stage, referral patterns, time to treatment, symptoms, and the proportion of patients assessed and supported by nursing and allied health professionals. Results: A total of 993 patients were included in the review, comprising 482 patients pre-WMCPC and 511 patients post-WMCPC. In the multivariate analysis, adjusting for ECOG (Eastern Cooperative Oncology Group) and stage, it was found that post-WMCPC patients experienced longer median overall survival (mOS, HR 0.84, 95% CI 0.72-0.98, p = 0.023). Furthermore, the time from referral to initial consultation date decreased significantly from 13.4 to 8.8 days in the post-WMCPC cohort (p < 0.001), along with a reduction in the time from the first clinic appointment to biopsy (14 vs. 8 days, p = 0.022). Additionally, patient-reported well-being scores showed improvement in the post-WMCPC cohort (p = 0.02), and these patients were more frequently attended to by nursing and allied health professionals (p < 0.001). Conclusions: The implementation of an inter-professional clinic for patients diagnosed with PDAC led to improvements in overall survival, patient-reported well-being, time to initial assessment visit and pathological diagnosis, and symptom management. These findings advocate for the adoption of an inter-professional clinic model in the treatment of patients with PDAC.

Genetic risk impacts the association of menopausal hormone therapy with colorectal cancer risk.

BACKGROUND: Menopausal hormone therapy (MHT), a common treatment to relieve symptoms of menopause, is associated with a lower risk of colorectal cancer (CRC). To inform CRC risk prediction and MHT risk-benefit assessment, we aimed to evaluate the joint association of a polygenic risk score (PRS) for CRC and MHT on CRC risk. METHODS: We used data from 28,486 postmenopausal women (11,519 cases and 16,967 controls) of European descent. A PRS based on 141 CRC-associated genetic variants was modeled as a categorical variable in quartiles. Multiplicative interaction between PRS and MHT use was evaluated using logistic regression. Additive interaction was measured using the relative excess risk due to interaction (RERI). 30-year cumulative risks of CRC for 50-year-old women according to MHT use and PRS were calculated. RESULTS: The reduction in odds ratios by MHT use was larger in women within the highest quartile of PRS compared to that in women within the lowest quartile of PRS (p-value = 2.7 × 10-8). At the highest quartile of PRS, the 30-year CRC risk was statistically significantly lower for women taking any MHT than for women not taking any MHT, 3.7% (3.3%-4.0%) vs 6.1% (5.7%-6.5%) (difference 2.4%, P-value = 1.83 × 10-14); these differences were also statistically significant but smaller in magnitude in the lowest PRS quartile, 1.6% (1.4%-1.8%) vs 2.2% (1.9%-2.4%) (difference 0.6%, P-value = 1.01 × 10-3), indicating 4 times greater reduction in absolute risk associated with any MHT use in the highest compared to the lowest quartile of genetic CRC risk. CONCLUSIONS: MHT use has a greater impact on the reduction of CRC risk for women at higher genetic risk. These findings have implications for the development of risk prediction models for CRC and potentially for the consideration of genetic information in the risk-benefit assessment of MHT use.

Cytidine deaminases APOBEC3C and APOBEC3D promote DNA replication stress resistance in pancreatic cancer cells.

Gemcitabine is a potent inhibitor of DNA replication and is a mainstay therapeutic for diverse cancers, particularly pancreatic ductal adenocarcinoma (PDAC). However, most tumors remain refractory to gemcitabine therapies. Here, to define the cancer cell response to gemcitabine, we performed genome-scale CRISPR-Cas9 chemical-genetic screens in PDAC cells and found selective loss of cell fitness upon disruption of the cytidine deaminases APOBEC3C and APOBEC3D. Following gemcitabine treatment, APOBEC3C and APOBEC3D promote DNA replication stress resistance and cell survival by deaminating cytidines in the nuclear genome to ensure DNA replication fork restart and repair in PDAC cells. We provide evidence that the chemical-genetic interaction between APOBEC3C or APOBEC3D and gemcitabine is absent in nontransformed cells but is recapitulated across different PDAC cell lines, in PDAC organoids and in PDAC xenografts. Thus, we uncover roles for APOBEC3C and APOBEC3D in DNA replication stress resistance and offer plausible targets for improving gemcitabine-based therapies for PDAC.

Body size and risk of colorectal cancer molecular defined subtypes and pathways: Mendelian randomization analyses.

BACKGROUND: Obesity has been positively associated with most molecular subtypes of colorectal cancer (CRC); however, the magnitude and the causality of these associations is uncertain. METHODS: We used Mendelian randomization (MR) to examine potential causal relationships between body size traits (body mass index [BMI], waist circumference, and body fat percentage) with risks of Jass classification types and individual subtypes of CRC (microsatellite instability [MSI] status, CpG island methylator phenotype [CIMP] status, BRAF and KRAS mutations). Summary data on tumour markers were obtained from two genetic consortia (CCFR, GECCO). FINDINGS: A 1-standard deviation (SD:5.1 kg/m2) increment in BMI levels was found to increase risks of Jass type 1MSI-high,CIMP-high,BRAF-mutated,KRAS-wildtype (odds ratio [OR]: 2.14, 95% confidence interval [CI]: 1.46, 3.13; p-value = 9 × 10-5) and Jass type 2non-MSI-high,CIMP-high,BRAF-mutated,KRAS-wildtype CRC (OR: 2.20, 95% CI: 1.26, 3.86; p-value = 0.005). The magnitude of these associations was stronger compared with Jass type 4non-MSI-high,CIMP-low/negative,BRAF-wildtype,KRAS-wildtype CRC (p-differences: 0.03 and 0.04, respectively). A 1-SD (SD:13.4 cm) increment in waist circumference increased risk of Jass type 3non-MSI-high,CIMP-low/negative,BRAF-wildtype,KRAS-mutated (OR 1.73, 95% CI: 1.34, 2.25; p-value = 9 × 10-5) that was stronger compared with Jass type 4 CRC (p-difference: 0.03). A higher body fat percentage (SD:8.5%) increased risk of Jass type 1 CRC (OR: 2.59, 95% CI: 1.49, 4.48; p-value = 0.001), which was greater than Jass type 4 CRC (p-difference: 0.03). INTERPRETATION: Body size was more strongly linked to the serrated (Jass types 1 and 2) and alternate (Jass type 3) pathways of colorectal carcinogenesis in comparison to the traditional pathway (Jass type 4). FUNDING: Cancer Research UK, National Institute for Health Research, Medical Research Council, National Institutes of Health, National Cancer Institute, American Institute for Cancer Research, Brigham and Women's Hospital, Prevent Cancer Foundation, Victorian Cancer Agency, Swedish Research Council, Swedish Cancer Society, Region Västerbotten, Knut and Alice Wallenberg Foundation, Lion's Cancer Research Foundation, Insamlingsstiftelsen, Umeå University. Full funding details are provided in acknowledgements.

Characterizing the burden of biliary tract cancers across 28 hospitals in Ontario, Canada.

BACKGROUND AND AIMS: The incidence of biliary tract cancers (BTC) appears to be increasing worldwide. We analyzed the characteristics of BTC-related hospitalizations under medical services across 28 hospitals in Ontario, Canada. METHODS: This study uses data collected by GEMINI, a hospital research data network. BTC-related hospitalizations from 2015 to 2021 under the Department of Medicine or intensive care unit were captured using the International Classification of Diseases, 10th revision, codes for intrahepatic cholangiocarcinoma (iCCA), extrahepatic cholangiocarcinoma, and gallbladder cancers. RESULTS: A total of 4596 BTC-related hospitalizations (2720 iCCA, 1269 extrahepatic cholangiocarcinoma, 607 gallbladder cancers) were analyzed. The number of unique patients with BTC-related hospitalizations increased over time. For iCCA-related hospitalizations, the total number of hospitalizations increased (from 385 in 2016 to 420 in 2021, p = .005), the hospital length of stay decreased over the study period (mean 10 days [SD, 12] in 2016 to 9 days [SD, 8] in 2021, p = .04), and the number of in-hospital deaths was stable (from 68 [18%] in 2016 to 55 [13%] in 2021, p = .62). Other outcomes such as 30-day readmissions, medical imaging tests, intensive care unit-specific hospitalizations, and length of stay were stable over time for all cohorts. The cost of hospitalization for the BTC cohort increased from median $8203 CAD (interquartile range, 5063-15,543) in 2017 to $8507 CAD (interquartile range, 5345-14,755) in 2021. CONCLUSIONS: This real-world data analysis showed a rising number of patients with BTC-related hospitalizations and rising number of iCCA-related hospitalizations across 28 hospitals in Ontario between 2015 and 2021.

The Mutographs biorepository: A unique genomic resource to study cancer around the world.

Large-scale biorepositories and databases are essential to generate equitable, effective, and sustainable advances in cancer prevention, early detection, cancer therapy, cancer care, and surveillance. The Mutographs project has created a large genomic dataset and biorepository of over 7,800 cancer cases from 30 countries across five continents with extensive demographic, lifestyle, environmental, and clinical information. Whole-genome sequencing is being finalized for over 4,000 cases, with the primary goal of understanding the causes of cancer at eight anatomic sites. Genomic, exposure, and clinical data will be publicly available through the International Cancer Genome Consortium Accelerating Research in Genomic Oncology platform. The Mutographs sample and metadata biorepository constitutes a legacy resource for new projects and collaborations aiming to increase our current research efforts in cancer genomic epidemiology globally.

Machine Learning to Allocate Palliative Care Consultations During Cancer Treatment.

PURPOSE: For patients with advanced cancer, early consultations with palliative care (PC) specialists reduce costs, improve quality of life, and prolong survival. However, capacity limitations prevent all patients from receiving PC shortly after diagnosis. We evaluated whether a prognostic machine learning system could promote early PC, given existing capacity. METHODS: Using population-level administrative data in Ontario, Canada, we assembled a cohort of patients with incurable cancer who received palliative-intent systemic therapy between July 1, 2014, and December 30, 2019. We developed a machine learning system that predicted death within 1 year of each treatment using demographics, cancer characteristics, treatments, symptoms, laboratory values, and history of acute care admissions. We trained the system in patients who started treatment before July 1, 2017, and evaluated the potential impact of the system on PC in subsequent patients. RESULTS: Among 560,210 treatments received by 54,628 patients, death occurred within 1 year of 45.2% of treatments. The machine learning system recommended the same number of PC consultations observed with usual care at the 60.0% 1-year risk of death, with a first-alarm positive predictive value of 69.7% and an outcome-level sensitivity of 74.9%. Compared with usual care, system-guided care could increase early PC by 8.5% overall (95% CI, 7.5 to 9.5; P < .001) and by 15.3% (95% CI, 13.9 to 16.6; P < .001) among patients who live 6 months beyond their first treatment, without requiring more PC consultations in total or substantially increasing PC among patients with a prognosis exceeding 2 years. CONCLUSION: Prognostic machine learning systems could increase early PC despite existing resource constraints. These results demonstrate an urgent need to deploy and evaluate prognostic systems in real-time clinical practice to increase access to early PC.

Genotoxic colibactin mutational signature in colorectal cancer is associated with clinicopathological features, specific genomic alterations and better survival.

Background and Aims: The microbiome has long been suspected of a role in colorectal cancer (CRC) tumorigenesis. The mutational signature SBS88 mechanistically links CRC development with the strain of Escherichia coli harboring the pks island that produces the genotoxin colibactin, but the genomic, pathological and survival characteristics associated with SBS88-positive tumors are unknown. Methods: SBS88-positive CRCs were identified from targeted sequencing data from 5,292 CRCs from 17 studies and tested for their association with clinico-pathological features, oncogenic pathways, genomic characteristics and survival. Results: In total, 7.5% (398/5,292) of the CRCs were SBS88-positive, of which 98.7% (392/398) were microsatellite stable/microsatellite instability low (MSS/MSI-L), compared with 80% (3916/4894) of SBS88 negative tumors (p=1.5x10-28). Analysis of MSS/MSI-L CRCs demonstrated that SBS88 positive CRCs were associated with the distal colon (OR=1.84, 95% CI=1.40-2.42, p=1x10-5) and rectum (OR=1.90, 95% CI=1.44-2.51, p=6x10-6) tumor sites compared with the proximal colon. The top seven recurrent somatic mutations associated with SBS88-positive CRCs demonstrated mutational contexts associated with colibactin-induced DNA damage, the strongest of which was the APC:c.835-8A>G mutation (OR=65.5, 95%CI=39.0-110.0, p=3x10-80). Large copy number alterations (CNAs) including CNA loss on 14q and gains on 13q, 16q and 20p were significantly enriched in SBS88-positive CRCs. SBS88-positive CRCs were associated with better CRC-specific survival (p=0.007; hazard ratio of 0.69, 95% CI=0.52-0.90) when stratified by age, sex, study, and by stage. Conclusion: SBS88-positivity, a biomarker of colibactin-induced DNA damage, can identify a novel subtype of CRC characterized by recurrent somatic mutations, copy number alterations and better survival. These findings provide new insights for treatment and prevention strategies for this subtype of CRC.

Clinically impactful metabolic subtypes of pancreatic ductal adenocarcinoma (PDAC).

Background: Pancreatic ductal adenocarcinoma (PDAC) is a lethal disease characterized by a diverse tumor microenvironment. The heterogeneous cellular composition of PDAC makes it challenging to study molecular features of tumor cells using extracts from bulk tumor. The metabolic features in tumor cells from clinical samples are poorly understood, and their impact on clinical outcomes are unknown. Our objective was to identify the metabolic features in the tumor compartment that are most clinically impactful. Methods: A computational deconvolution approach using the DeMixT algorithm was applied to bulk RNASeq data from The Cancer Genome Atlas to determine the proportion of each gene's expression that was attributable to the tumor compartment. A machine learning algorithm designed to identify features most closely associated with survival outcomes was used to identify the most clinically impactful metabolic genes. Results: Two metabolic subtypes (M1 and M2) were identified, based on the pattern of expression of the 26 most important metabolic genes. The M2 phenotype had a significantly worse survival, which was replicated in three external PDAC cohorts. This PDAC subtype was characterized by net glycogen catabolism, accelerated glycolysis, and increased proliferation and cellular migration. Single cell data demonstrated substantial intercellular heterogeneity in the metabolic features that typified this aggressive phenotype. Conclusion: By focusing on features within the tumor compartment, two novel and clinically impactful metabolic subtypes of PDAC were identified. Our study emphasizes the challenges of defining tumor phenotypes in the face of the significant intratumoral heterogeneity that typifies PDAC. Further studies are required to understand the microenvironmental factors that drive the appearance of the metabolic features characteristic of the aggressive M2 PDAC phenotype.

Analysis of prognostic germline polymorphisms in patients with advanced hepatocellular carcinoma.

Background: The prognosis of hepatocellular carcinoma (HCC) is influenced by both tumor and patient specific factors. Current therapies of advanced HCC target angiogenesis and immune evasion, however there are no clinically useful biomarkers to guide clinicians. Methods: Our aim in this retrospective cohort study was to validate single nucleotide polymorphisms (SNPs) prognostic of outcome in advanced HCC from the literature, and to analyze exploratory SNPs chosen from evaluation of the HCC tumor immune microenvironment. Using a database of patients with HCC treated with sorafenib, blood samples were genotyped, clinical variables were retrospectively collected, and SNPs were analyzed for association with progression-free survival (PFS) and overall survival (OS). A subsequent analysis was conducted to determine if identified SNPs were prognostic in trans arterial chemoembolization (TACE) treated patients. Results: Literature review identified 7 SNPs in vascular endothelial growth factor (VEGF), eNOS, angiopoietin 2 (ANGPT2) and vascular endothelial growth factor receptor 2 (VEGFR2), however none were externally validated in our dataset. Of the 35 exploratory immunomodulatory SNPs, the following were associated with PFS or OS: CCL2 C-C motif ligand 2 (CCL2) (rs1024611), interleukin-10 (IL-10) (rs1800896), cytotoxic T-lymphocyte antigen-4 (CTLA-4) (rs231775) and NFKB1 (rs28362491). Conclusions: SNPs identified by literature review to be prognostic in sorafenib treated patients with advanced HCC were not validated in our dataset. Our findings suggest potentially important prognostic implications of SNPs in VEGFR2, CCL2, IL-10, CTLA-4 and NFKB1 that deserve further study.

Deciphering the Pathways to PARP Sensitivity in Pancreatic Cancer.

A recent article analyzed paired cell-free DNA among patients with platinum-sensitive BRCA- or PALB2-mutated pancreatic cancer who received maintenance olaparib. Reversion mutations were linked with worse outcomes. These types of paired correlative studies are needed to improve our understanding of drug resistance and guide future clinical trials. See related article by Brown et al., p. 5207.

Erlotinib Plus Gemcitabine Compared With Gemcitabine Alone in Patients With Advanced Pancreatic Cancer: A Phase III Trial of the National Cancer Institute of Canada Clinical Trials Group.

PURPOSE: Patients with advanced pancreatic cancer have a poor prognosis and there have been no improvements in survival since the introduction of gemcitabine in 1996. Pancreatic tumors often overexpress human epidermal growth factor receptor type 1 (HER1/EGFR) and this is associated with a worse prognosis. We studied the effects of adding the HER1/EGFR-targeted agent erlotinib to gemcitabine in patients with unresectable, locally advanced, or metastatic pancreatic cancer. PATIENTS AND METHODS: Patients were randomly assigned 1:1 to receive standard gemcitabine plus erlotinib (100 or 150 mg/d orally) or gemcitabine plus placebo in a double-blind, international phase III trial. The primary end point was overall survival. RESULTS: A total of 569 patients were randomly assigned. Overall survival based on an intent-to-treat analysis was significantly prolonged on the erlotinib/gemcitabine arm with a hazard ratio (HR) of 0.82 (95% CI, 0.69 to 0.99; P = .038, adjusted for stratification factors; median 6.24 months v 5.91 months). One-year survival was also greater with erlotinib plus gemcitabine (23% v 17%; P = .023). Progression-free survival was significantly longer with erlotinib plus gemcitabine with an estimated HR of 0.77 (95% CI, 0.64 to 0.92; P = .004). Objective response rates were not significantly different between the arms, although more patients on erlotinib had disease stabilization. There was a higher incidence of some adverse events with erlotinib plus gemcitabine, but most were grade 1 or 2. CONCLUSION: To our knowledge, this randomized phase III trial is the first to demonstrate statistically significantly improved survival in advanced pancreatic cancer by adding any agent to gemcitabine. The recommended dose of erlotinib with gemcitabine for this indication is 100 mg/d.

Machine Learning-Based Early Warning Systems for Acute Care Utilization During Systemic Therapy for Cancer.

BACKGROUND: Emergency department visits and hospitalizations frequently occur during systemic therapy for cancer. We developed and evaluated a longitudinal warning system for acute care use. METHODS: Using a retrospective population-based cohort of patients who started intravenous systemic therapy for nonhematologic cancers between July 1, 2014, and June 30, 2020, we randomly separated patients into cohorts for model training, hyperparameter tuning and model selection, and system testing. Predictive features included static features, such as demographics, cancer type, and treatment regimens, and dynamic features, such as patient-reported symptoms and laboratory values. The longitudinal warning system predicted the probability of acute care utilization within 30 days after each treatment session. Machine learning systems were developed in the training and tuning cohorts and evaluated in the testing cohort. Sensitivity analyses considered feature importance, other acute care endpoints, and performance within subgroups. RESULTS: The cohort included 105,129 patients who received 1,216,385 treatment sessions. Acute care followed 182,444 (15.0%) treatments within 30 days. The ensemble model achieved an area under the receiver operating characteristic curve of 0.742 (95% CI, 0.739-0.745) and was well calibrated in the test cohort. Important predictive features included prior acute care use, treatment regimen, and laboratory tests. If the system was set to alarm approximately once every 15 treatments, 25.5% of acute care events would be preceded by an alarm, and 47.4% of patients would experience acute care after an alarm. The system underestimated risk for some treatment regimens and potentially underserved populations such as females and non-English speakers. CONCLUSIONS: Machine learning warning systems can detect patients at risk for acute care utilization, which can aid in preventive intervention and facilitate tailored treatment. Future research should address potential biases and prospectively evaluate impact after system deployment.

Performance evaluation of a North American center using the established global benchmark for laparoscopic liver resections: A retrospective study.

BACKGROUND: The global benchmark cut-offs were set for laparoscopic liver resection procedures: left lateral sectionectomy, left hepatectomy, and right hepatectomy. We aimed to compare the performance of our North American center with the established global benchmarks. METHODS: This is a single-center study of adults who underwent laparoscopic liver resection between 2010 to 2022 at the Toronto General Hospital. Fourteen benchmarking outcomes were assessed: operation time, intraoperative blood transfusion, estimated blood loss, blood loss ≥500 mL, blood loss ≥1000mL, open-conversion, postoperative length of stay, return to operation, postoperative morbidity, postoperative major-morbidity, 30-day mortality, 90-day mortality, R1 resection, and failure to rescue. Low-risk benchmark cases were defined as follows: patients aged 18 to 70 years, American Society of Anesthesiologist score ≤ 2, tumor size <10 cm, and Child-Pugh score ≤A. Cases involving bilio-enteric anastomosis, hilar dissection, or concomitant major procedures were excluded from the low-risk category. Cases that did not meet the criteria for low-risk selection were considered high-risk cases. RESULTS: A total of 178 laparoscopic liver resection cases were analyzed (109 left lateral sectionectomies, 45 left hepatectomies, 24 right hepatectomies). Forty-four (25%) cases qualified as low-risk cases (23 left lateral sectionectomies, 16 left hepatectomies, 5 right hepatectomies). The postoperative major morbidity and 90-day mortality after left lateral sectionectomy, left hepatectomy, and right hepatectomy for the low-risk cases were 0%, 0%, and 0%, and 0%, 0%, and 0%, respectively. For the high-risk cases post-2017, the outcomes in the same order were 0%, 0%, and 12%; 0%, 0%, and 0%, respectively. For the high-risk cases operated pre2017, the outcomes in the same order were 9%∗, 16%∗, and 18%; 2%∗, 0%, and 9%∗ (asterisks indicate not meeting the global cut-off), respectively. CONCLUSION: A North American center was able to achieve outcomes comparable to the established global benchmark for laparoscopic liver resection.