Badat, Saad

Purpose This study aimed to develop and validate a tool to predict the primary tumor site in patients with metastatic cancer of unknown primary using demographic and clinical variables, including imaging-derived sites of metastasis, without pathology or laboratory data. Methods In this population-based study, data were extracted from the SEER database. Patients diagnosed between 2010 and 2020 with a known primary cancer site, metastatic staging, and metastasis in the bone, liver, brain, and lungs were included. Patients were randomly split into training (70%) and/or testing (30%) sets. A total of 27 primary tumor sites were included. A gradient boosting model was trained to predict the primary tumor site, with random forest and classification tree models evaluated as comparators. Random guess and frequency-weighted guessing models were used for baseline comparison. Results A total of 689,715 patients were included from 9,208,295 queried. The most common primary cancer sites were non–small cell lung cancer (NSCLC) (n = 211,338), intestines (n = 64,187), and pancreas (n = 60,310). The baseline models achieved Top-5 accuracies of 0.179 for a random guess and 0.476 for a frequency-weighted guess. The gradient boosting model achieved Top-1, Top-3, and Top-5 accuracies of 0.40 ± 0.001, 0.680 ± 0.001, and 0.790 ± 0.001, respectively, across 100 Monte-Carlo replicates. Top-5 accuracy was highest for primary NSCLC (0.990 ± 0.001) and intestinal cancer (0.900 ± 0.003). Our model was most influenced by sex (gain = 0.27) and metastasis to bone (gain = 0.20). A web app was constructed: primarycancerpredictor.netlify.app. Conclusion This study developed and evaluated a gradient boosting machine learning model and a user-friendly web tool to aid in the diagnosis of cancer of unknown primary.

PURPOSE Although outcomes for nonmetastatic cancers have improved since the 1970s, it remains uncertain whether long-term survival has improved among patients presenting with de novo metastatic disease, defined as cancer presenting with distant metastases at the time of initial diagnosis. This study evaluated changes in 5-year survival among patients with de novo metastatic cancer diagnosed in 1976 versus 2015, with a minimum 5-year follow-up through 1981 and 2020, respectively. METHODS Population-based data were obtained from the SEER database. Patients diagnosed with de novo metastatic cancer in 1976 (n = 11,864) and 2015 (n = 26,324) were included. Five-year overall survival (OS) was compared between cohorts, and multivariable logistic regression identified demographic and clinical factors associated with survival. RESULTS The 5-year OS increased significantly from 10.6% in 1976-1981 to 23.6% in 2015-2020 ( P < .001). Survival gains were largest in breast cancer (12.6%-34.2%; P < .001), ovarian cancer (14.4%-32.5%; P < .001), colorectal cancer (3.7%-15.0%; P < .001), and hematologic malignancies (24.3%-53.9%; P < .001). Minimal improvement occurred in pancreatic cancer (0.25%-1.6%; P = .007) and small cell lung cancer (0.8%-3.1%; P = .008). Multivariable analysis demonstrated that older age, male sex, and Black race were associated with lower odds of 5-year survival (all P < .01) CONCLUSION Over the past four decades, long-term survival among patients with de novo metastatic cancer has more than doubled; however, progress remains uneven across cancer types. Gains appear concentrated in malignancies with substantial therapeutic advances, whereas others show persistently poor outcomes. Continued efforts are needed to address disparities and improve survival in resistant cancer types.