Background: Although cancer patients are increasingly admitted to the intensive care unit (ICU) for cancer- or treatment-related complications, improved mortality prediction remains a big challenge. This study describes a new ML-based mortality prediction model for critically ill cancer patients admitted to ICU. Patients and Methods: We developed CanICU, a machine learning-based 28-day mortality prediction model for adult cancer patients admitted to ICU from Medical Information Mart for Intensive Care (MIMIC) database in the USA (n = 766), Yonsei Cancer Center (YCC, n = 3571), and Samsung Medical Center in Korea (SMC, n = 2563) from 2 January 2008 to 31 December 2017. The accuracy of CanICU was measured using sensitivity, specificity, and area under the receiver operating curve (AUROC). Results: A total of 6900 patients were included, with a 28-day mortality of 10.2%/12.7%/36.6% and a 1-year mortality of 30.0%/36.6%/58.5% in the YCC, SMC, and MIMIC-III cohort. Nine clinical and laboratory factors were used to construct the classifier using a random forest machine-learning algorithm. CanICU had 96% sensitivity/73% specificity with the area under the receiver operating characteristic (AUROC) of 0.94 for 28-day, showing better performance than current prognostic models, including the Acute Physiology and Chronic Health Evaluation (APACHE) or Sequential Organ Failure Assessment (SOFA) score. Application of CanICU in two external data sets across the countries yielded 79–89% sensitivity, 58–59% specificity, and 0.75–0.78 AUROC for 28-day mortality. The CanICU score was also correlated with one-year mortality with 88–93% specificity. Conclusion: CanICU offers improved performance for predicting mortality in critically ill cancer patients admitted to ICU. A user-friendly online implementation is available and should be valuable for better mortality risk stratification to allocate ICU care for cancer patients.

Background: Programmed death‐ligand 1 (PD‐L1) is an important screening biomarker to select patients with gastric cancer (GC) for optimized treatment, including immune checkpoint inhibitors (ICI).Methods: In this single‐institution retrospective cohort study, patients with metastatic GC with available PD‐L1 results between October 2019 and September 2021 were identified by reviewing their electronic medical records. Genomic data were obtained from the Samsung Medical Center Clinical Sequencing Platform.Results: Among the 399 patients, 276 (69%) had a PD‐L1 combined positive score (CPS) ≥1, 155 (39%) had a CPS between 1 and 5, and 121 (30%) had a CPS ≥5. Of the 121 patients with CPS ≥5, 28 (23%) had a known etiology for “inflamed tumor,” with Epstein–Barr virus (EBV) positivity (N = 11) or high tumor mutational burden (TMB) (N = 17), which included microsatellite instability (MSI) (N = 9). PD‐L1 CPS ≥5 was observed in 11/11 (100%) patients with EBV positivity, 9/12 (75%) patients with MSI, and 17/33 (52%) patients with high TMB. For the 108 patients who received ICI therapy, CPS ≥5 was the only predictor significantly associated with survival in multivariable analyses, including TMB, MSI, or EBV. Objective response rate (ORR) was 49% in patients with CPS ≥5, 30% in patients with 1 ≤ CPS <5, and 19% in patients with CPS <1. Among the 31 responders to ICI therapy, 27 (87%) had a CPS of ≥1. Mutations in TET2, IRS2, DOT1L, PTPRT, and LRP1B were associated with a higher ORR (63%–100%), whereas MDC1 mutations were associated with a low ORR (22%).Conclusions: PD‐L1 expression is an independent and sensitive biomarker for ICI therapy. Considering its significant association with several gene alterations, including PIK3CA mutations and MET amplification, combining ICI therapy with other targeted agents may be a promising therapeutic strategy for GC.

With the advances in chemotherapy and immunotherapy, a small subset of patients may be eligible for conversion surgery after achieving tumor regression with chemotherapy. This is a retrospective cohort study of 118 patients with stage IV gastric cancer who received palliative chemotherapy and conversion surgery with a negative resection margin at Samsung Medical Center. Baseline features included comorbidities, body mass index (BMI), carcinoembryonic antigen (CEA) level, primary tumor size, biopsy histology, distant metastatic sites, and molecular markers—HER2, MSI/MMR, PD-L1, and EBV. Post-chemotherapy features included BMI, CEA level, chemotherapy regimen, objective response to chemotherapy, and number of preoperative chemotherapy cycles. Post-operational features included tumor size, histologic differentiation and Lauren’s classification, pathologic tumor and nodal stages, invasion of lymphatics/vessels/nerves, peritoneal cytology, and the receipt of postoperative chemotherapy. Of 118 patients, 60 patients received total gastrectomy and 58 patients received subtotal gastrectomy. In all, 21 patients achieved a pathologic complete response, and 97 patients achieved downstaging to yp stage I, II, or III. Before conversion surgery, patients received first-line capecitabine/oxaliplatin (62%), HER2 inhibitors combined with chemotherapy (18%), immune checkpoint inhibitors (15%), and inhibitors of MET or VEGFR2 (5%). In the multivariable analysis, BMI at the time of diagnosis, either HER2 positive, high MSI, or deficient MMR, and the use of targeted agents were significant prognostic factors. Conversion surgery could be considered in patients with stage IV gastric cancer regardless of the initial disease burden. BMI and molecular markers are important prognostic factors that can be used to select candidates.

Click on the article title to read more.

Advancements in the handling of comprehensive genetic data in cancer research have led to the expansion of molecular subtyping studies. These studies reflect not only conventional tumor biological prognostic factors but also strive to develop predictive testing for therapeutic responses. While significant progress has been achieved, with commercial-grade assays now routinely used in breast cancer, similar efforts are currently underway in gastric cancer. In this review, we shed light on the current consensus in molecular subtyping research in gastric cancer and explore the potential of identified molecular signatures for the development of prognostic and predictive testing. Additionally, we address the unique characteristics of gastric cancer that present challenges for the straightforward development of successful prognostic/predictive tests. Drawing from these insights, we provide recommendations for incorporating prognostic testing into clinical treatment options and highlight key considerations for the successful advancement of predictive testing research.

Shin MK, et al. Foregut Surg. 2023 Jul;3(2):40-48. https://doi.org/10.51666/fs.2023.3.e8