JCO precision oncology最新文献

Purpose: To develop and validate machine learning (ML) models for postoperative risk stratification in oral cavity squamous cell carcinoma (OCSCC) and to examine whether ML-derived risk groups modify the association between adjuvant therapy and overall survival (OS).

Methods: Using the National Cancer Database, we identified adults with invasive OCSCC treated with primary surgery. The surgery-alone cohort (n = 18,543) was split 70/30 for training/testing to develop DeepSurv, Neural Multi-Task Logistic Regression (NMTLR), and Random Survival Forest (RSF) models. Risk scores were generated for the full cohort (n = 35,625) and converted to low, intermediate, and high groups. Within groups, treatment effects of adjuvant radiotherapy (RT) and chemoradiotherapy (CRT) were estimated using multivariable Cox models.

Results: The best performance was achieved by DeepSurv (C-index 0.73), with similar discrimination for NMTLR/RSF (C-index 0.71-0.72). For DeepSurv, the full cohort was partitioned into low- (50.0%), intermediate- (32.0%), and high-risk (18.0%) groups with distinct 5-year OS rates: 77.6%, 53.0%, and 29.3%, respectively. In the low-risk group, adjuvant RT (adjusted hazard ratios [aHR], 0.94 [95% CI, 0.87 to 1.02]) and CRT (aHR, 1.03 [95% CI, 0.91 to 1.17]) did not improve OS. In the intermediate-risk group, OS improved with RT (aHR, 0.61 [95% CI, 0.57 to 0.65]) and CRT (aHR, 0.56 [95% CI, 0.52 to 0.61]). In the high-risk group, both adjuvant RT (aHR, 0.47 [95% CI, 0.43 to 0.51]) and CRT (aHR, 0.39 [95% CI, 0.36 to 0.41]) were associated with improved OS compared with surgery alone. CRT was associated with a modest benefit compared with RT. NMTLR and RSF yielded concordant patterns. Top features included pT4a stage, age ≥70 years, and extranodal extension.

Conclusion: ML-derived risk stratification identifies patients with OCSCC most likely to benefit from adjuvant therapy, supporting intensification for intermediate-/high-risk patients and potential deintensification for low-risk patients. External prospective validation is warranted to enable clinical implementation.

Purpose: Although we typically consider diffuse pleural mesothelioma (DPM) as a disease of elderly males with a history of asbestos exposure, young individuals are also affected. Given the increasing incidence of cancer in young individuals, we investigated the characteristics and underlying biology of young patients with DPM.

Methods: We identified patients with DPM age ≤50 years from Memorial Sloan Kettering Cancer Center (1990-2023). Clinicopathologic and genetic characteristics were compared between very young (≤35 years) and young (36-50 years) patients.

Results: We identified 273 patients with DPM age ≤50 years; 39% were female, 40% had a history of self-reported occupational asbestos exposure (very young, 16% v young, 45%; P < .01), 16% had a personal history of cancer (very young, 7% v young, 18%; P = .07), and 70% had a family history of cancer (most commonly breast, lung, and colon). Most frequently altered somatic genes (51/273 tested; 46/51 positive) included BAP1 (22/47, 47%), NF2(18/47, 38%), CDKN2A/B (12/48, 25%), and TP53 (10/48, 21%). Very young patients had less somatic alterations in BAP1 (13% v 63%, P < .01) and a lower tumor mutational burden (0.9 v 1.8 mut/mB, P < .01); they were less frequently carriers of germline alterations (0% [0/8] v 43% [6/14], P = .05). Somatic alterations in CDKN2B were more frequent among females (females 42% v males 12%, P = .03).

Conclusion: Young patients with DPM have strong personal and family histories of cancer, and heterogeneous somatic and germline alterations, indicating divergent underlying biology. With the increasing prevalence of young adults with cancers, mesotheliomas, although uncommon, should be on the differential for patients even without asbestos exposure history in this age group.

Purpose: Recurrent chromosomal fusions in neurotrophic tropomyosin receptor kinase (NTRK) have been reported as an oncogenic driver across human cancer. Herein, we report results from subprotocol Z1E of the NCI-MATCH (Molecular Analysis for Therapy Choice) trial investigating the efficacy and safety of larotrectinib, a highly selective inhibitor of all three TRK proteins, in TRK fusion-positive cancers.

Methods: From August 21, 2017, to July 8, 2021, patients with solid tumors or lymphomas progressing on standard therapy were genomically profiled and enrolled to NCI-MATCH, a genomically driven, signal-seeking, precision medicine platform trial. They were assigned to subprotocol Z1E if a NTRK fusion-positive tumor was identified. Patients received larotrectinib 100 mg twice daily until disease progression or unacceptable toxicity. The primary end point was objective response rate (ORR). Secondary end points included progression-free survival (PFS) at 6 months, PFS, overall survival (OS), and safety.

Results: In total, 12 patients with centrally confirmed NTRK fusion-positive tumors were treated with larotrectinib across six distinct tumor histologies. The ORR was 75% (90% CI, 47.3% to 92.8%) with median PFS and OS of 14.4 (90% CI, 8.48 to not reached [NR]) months and 23.9 (90% CI, 11.1 to NR) months, respectively. Treatment was well tolerated, with adverse events predominantly being grade 1.

Conclusion: NCI-MATCH subprotocol Z1E demonstrates that larotrectinib offers clinically significant benefit with marked ORR across NTRK fusion-positive tumors with no clinically significant toxicities. These findings support the rationale for the US Food and Drug Administration's tissue-agnostic approval of larotrectinib for NTRK fusion-positive tumors. Our findings demonstrate the necessity of including NTRK1, NTRK2, and NTRK3 within comprehensive molecular assays of cancer to navigate patients to an easily administered and highly effective therapy.

Purpose: This study aimed to characterize the genomic landscape of Korean gastric cancer and evaluate associations among oncogenic alterations, established biomarkers, demographics, and treatment outcomes.

Methods: A total of 1,283 patients with gastric cancer who underwent tumor-only targeted sequencing as part of practice and received palliative treatment between January 2017 and August 2025 at the Samsung Medical Center were included.

Results: Among 1,283 patients (median [IQR] age, 61 [52-68] years; 827 males [64.46%]), TP53 (51.91%), ARID1A (19.02%), ERBB2 (12%), KRAS (10.29%), and PIK3CA (9.12%) were the most frequently altered genes. Epstein-Barr virus-positive tumors exhibited enrichment of BCOR, PIK3CA, and ARID1A alterations and reduced TP53 mutations (false discovery rate [FDR] adjusted P < .01). Human epidermal growth factor receptor 2-positive tumors were characterized by coamplification of ERBB2, CCNE1, and MYC (FDR adjusted P < .001), whereas PD-L1 positivity was associated with KRAS and CDKN2A alterations (FDR-adjusted P < .05). Among patients treated with first-line nivolumab plus chemotherapy (n = 269), those with high tumor mutational burden (TMB; ≥10 mutations per megabase) had improved overall survival (v the low TMB subgroup; hazard ratio [HR], 0.48 [95% CI, 0.25 to 0.93]; P = .03), particularly when combined with PD-L1 positivity (v all other biomarker-defined subgroups; HR, 0.33 [95% CI, 0.14 to 0.76]; P = .006). Moreover, as TMB levels increased, patients derived greater survival benefit from nivolumab plus chemotherapy versus chemotherapy alone, even among those with microsatellite-stable tumors. Across treatment regimens, FGFR2 and MET alterations were linked to poorer outcomes, whereas PIK3CA mutations were observed in patients with longer overall survival after first-line chemotherapy.

Conclusion: Our findings provide a comprehensive genomic landscape of Korean gastric cancer and underscore the clinical relevance of integrating genomic and established biomarkers to advance precision oncology.

Purpose: To build a self-supervised magnetic resonance imaging (MRI) foundation model from routine clinical scans and to test whether it can support key glioma-related applications, including post-therapy imaging outcome characterization and molecular marker inference.

Materials and methods: We created the Unified Multimodal Brain Imaging Foundation (UMBIF) model and pretrained it in a self-supervised manner using 51,029 routine brain MRI examinations collected across multiple institutions. Pretraining used a hybrid objective that couples masked-image reconstruction with contrastive representation learning to encourage anatomically and clinically informative embeddings. The pretrained UMBIF encoder was then adapted to downstream multicenter data sets to predict (1) post-treatment radiographic outcomes and (2) molecular biomarkers, including IDH mutation, MGMT promoter methylation, and 1p/19q codeletion. Performance was benchmarked against commonly used convolutional networks and traditional machine learning classifiers, using accuracy, sensitivity, specificity, and receiver operating characteristic-AUC as primary metrics.

Results: Relative to self-supervised initialization derived from natural-image corpora or from approaches emphasizing only large tumor-area crops (self-supervised learning [SSL]-ImageNet and SSL-Cerebral), the UMBIF encoder-decoder design captured richer, more task-relevant features and consistently improved downstream discrimination. The best pretrained model achieved an accuracy of 0.899 (AUC, 0.815) for post-treatment radiographic outcome characterization. For molecular profiling, it reached accuracies/AUCs of 0.898/0.916 for 1p/19q codeletion, 0.829/0.896 for IDH mutation status, and 0.905/0.859 for MGMT promoter methylation, indicating strong potential utility in clinical decision support.

Conclusion: UMBIF showed robust transferability to both post-therapy imaging assessment and molecular status prediction in glioma. By leveraging large-scale self-supervised pretraining to boost performance while reducing dependence on manual annotations, the framework may facilitate more efficient and reliable diagnostic workflows.

Purpose: Multifactorial breast cancer (BC) risk prediction models use a range of predictors to estimate an individual's chance of developing BC. Data on risk factors are often incomplete, and point estimates calculated when data are missing can mask considerable uncertainty. Quantifying this uncertainty is critical for effective risk communication.

Methods: We used Monte Carlo simulation methods to estimate the distribution of 10-year BC risk for individuals with missing data, using the BOADICEA multifactorial model as an example. Multivariate imputation by chained equations with large representative reference data sets was used to sample missing covariates. We developed a framework for estimating the uncertainty distribution, uncertainty intervals (UIs), and probability of reclassification, which can be applied to any given individual with missing risk factor data. This was applied to estimating individual-level uncertainty distributions and quantifying the probability of reclassification when groups of risk factors are measured, for a range of example women.

Results: Women with limited risk factor data had considerable uncertainty in their estimated BC risk, and 95% UIs spanned all risk categories. This was especially relevant for women classified as moderate-risk, such as those with strong family history or a moderate-risk pathogenic variant. Reclassification probability in this case was as high as 57.5%, with 95% UI of 0.9% to 9.3% for the 10-year risk from age 40 years. Risk certainty improved with additional data collection, particularly genetic information or mammographic density measurement.

Conclusion: Our results demonstrate that, in some cases, there is considerable probability of reclassification after collecting missing data. Methodology presented here can identify situations where it would be most beneficial to collect additional information, to enable better informed clinical decision making.

Purpose: The role of circulating tumor DNA (ctDNA) monitoring in predicting treatment response and survival outcomes in early-stage and metastatic solid tumors has shown promising results. The value of early ctDNA dynamics in solid tumors treated with immune checkpoint inhibitors (ICIs) merits further investigation. Our study aims to assess the role of early ctDNA changes, measured 3-4 weeks after initiating ICI therapy, in predicting clinical outcomes in patients with advanced-stage melanoma.

Methods: We performed a multi-institutional analysis of real-world data from testing patients with unresectable stage III/IV melanoma using a personalized, tumor-informed ctDNA assay (Signatera) during the course of treatment with anti-PD-1-based therapy. ctDNA levels were assessed before the start of treatment and at 3-4 weeks before the second treatment dose. The change in ctDNA (decrease v increase) between the two points was assessed to evaluate the odds of objective response and disease control and to evaluate the risk of progression-free survival (PFS) and overall survival (OS).

Results: One hundred seventeen patients were evaluated. Regression analysis revealed that a decrease in ctDNA was associated with increased odds of disease control (odds ratio [OR], 30.56 [95% CI, 10.64 to 87.84], P < .001) and objective response (OR, 23.54 [95% CI, 8.58 to 64.57], P < .001). Patients with a decrease in ctDNA had an increased PFS (hazard ratio [HR], 0.18 [95% CI, 0.11 to 0.31], P < .001) and OS (HR, 0.28 [95% CI, 0.13 to 0.56], P < .001) probability.

Conclusion: We found that early ctDNA dynamics after only 3-4 weeks of ICI initiation in patients with advanced-stage melanoma appears to be a candidate strategy to predict overall treatment response, risk of progression, and long-term survival. Larger prospective studies are warranted to validate the utility of early ctDNA changes in treatment monitoring.