Machine Learning-Assisted Recurrence Prediction for Patients With Early-Stage Non-Small-Cell Lung Cancer.

PURPOSE: Stratifying patients with cancer according to risk of relapse can personalize their care. In this work, we provide an answer to the following research question: How to use machine learning to estimate probability of relapse in patients with early-stage non-small-cell lung cancer (NSCLC)? MATERIALS AND METHODS: For predicting relapse in 1,387 patients with early-stage (I-II) NSCLC from the Spanish Lung Cancer Group data (average age 65.7 years, female 24.8%, male 75.2%), we train tabular and graph machine learning models. We generate automatic explanations for the predictions of such models. For models trained on tabular data, we adopt SHapley Additive exPlanations local explanations to gauge how each patient feature contributes to the predicted outcome. We explain graph machine learning predictions with an example-based method that highlights influential past patients. RESULTS: Machine learning models trained on tabular data exhibit a 76% accuracy for the random forest model at predicting relapse evaluated with a 10-fold cross-validation (the model was trained 10 times with different independent sets of patients in test, train, and validation sets, and the reported metrics are averaged over these 10 test sets). Graph machine learning reaches 68% accuracy over a held-out test set of 200 patients, calibrated on a held-out set of 100 patients. CONCLUSION: Our results show that machine learning models trained on tabular and graph data can enable objective, personalized, and reproducible prediction of relapse and, therefore, disease outcome in patients with early-stage NSCLC. With further prospective and multisite validation, and additional radiological and molecular data, this prognostic model could potentially serve as a predictive decision support tool for deciding the use of adjuvant treatments in early-stage lung cancer.

Seroprevalence and immunological memory against SARS-CoV-2 in lung cancer patients: the SOLID study.

BACKGROUND: At present, we did not find any articles that studied seroprevalence and its persistence several months later in lung cancer patients in the setting of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. Most patients with coronavirus disease 2019 (COVID-19) go on to develop antibodies (Abs) against viral proteins. However, it is not known how long these Abs last nor whether cancer treatments could affect the duration of immune response. METHODS: This prospective, longitudinal, multicenter serological study in the setting of SARS-CoV-2 infection was carried out in 50 Spanish hospitals. Eligibility criterion was the diagnosis of any lung cancer. The determination of anti-SARS-CoV-2 IgG Abs was performed by qualitative immuno-enzymatic assay using enzyme-linked immunosorbent assay (ELISA) kit from NovaLisa whose Abs target the recombinant antigen N of the nucleocapsid of SARS-CoV-2. The first Ab determination was performed between April 21 and June 3, 2020. The second Ab determination was performed in all previously seropositive patients, between September 10 and November 20, 2020. Study objectives were to prospectively determine seroprevalence in unselected lung cancer patients during the first wave of the pandemic; the persistence of immunity; protection or lack thereof against reinfection; and the influence of treatments on maintenance or loss of immunity. RESULTS: Of 1,500 patients, 128 were seropositive, overall prevalence of 8.5% seropositivity [95% confidence interval (CI): 7.2-10.1%]. Seventy-five percent were in active cancer treatment. Forty-seven point seven percent of IgG positive participants had experienced a symptomatic illness suspected of being infected with SARS-CoV-2 (95% CI: 38.8-56.6%). A second determination was performed on average 4.5 months later [interquartile range (IQR), 4.0-5.0 months] and obtained for 104 of the initially seropositive patients (81%), it could not be obtained in 24 patients, the majority due to death caused by disease progression (73%). In the second determination, IgG was not detected in 30.8% of patients. The severity of the infection, the need for hospitalization (P=0.032) and the presence of symptoms at diagnosis (P=0.02) were associated with persistence of immunity in the second determination. No variables or treatments received were associated with Abs loss. CONCLUSIONS: Immunity against SARS-CoV-2 does not appear to be compromised by treatment and persists beyond 4 months. Neither do mortality rates appear to be particularly high in this unselected population. TRIAL REGISTRATION: ClinicalTrials.gov identifier: NCT04407143.

Epidemiology, treatment, and survival in small cell lung cancer in Spain: Data from the Thoracic Tumor Registry.

BACKGROUND: Small-cell lung cancer (SCLC) is an aggressive disease with high metastatic potential and poor prognosis. Due to its low prevalence, epidemiological and clinical information of SCLC patients retrieved from lung cancer registries is scarce. PATIENTS AND METHODS: This was an observational multicenter study that enrolled patients with lung cancer and thoracic tumors, recruited from August 2016 to January 2020 at 50 Spanish hospitals. Demographic and clinical data, treatment patterns and survival of SCLC patients included in the Thoracic Tumor Registry (TTR) were analyzed. RESULTS: With a total of 956 cases, the age of 64.7 ± 9.1 years, 78.6% were men, 60.6% smokers, and ECOG PS 0, 1 or ≥ 2 in 23.1%, 53.0% and 23.8% of cases, respectively. Twenty percent of patients had brain metastases at the diagnosis. First-line chemotherapy (CT), mainly carboplatin or cisplatin plus etoposide was administered to >90% of patients. In total, 36.0% and 13.8% of patients received a second and third line of CT, respectively. Median overall survival was 9.5 months (95% CI 8.8-10.2 months), with an estimated rate of 70.3% (95% CI 67.2-73.4%), 38.9% (95% CI 35.4-42.4%), and 14.8% (95% CI 11.8-17.8%) at 6, 12 and 24 months respectively. Median progression-free survival was 6.3 months. Higher mortality and progression rates were significantly associated with male sex, older age, smoking habit, and ECOG PS 1-2. Long-term survival (> 2 years) was confirmed in 6.6% of patients, showing a positive correlation with better ECOG PS, poor smoking and absence of certain metastases at diagnosis. CONCLUSION: This study provides an updated overview of the clinical situation and treatment landscape of ES-SCLC in Spain. Our results might assist oncologists to improve current clinical practice towards a better prognosis for these patients.

Analysis of circulating tumour DNA to identify patients with epidermal growth factor receptor-positive non-small cell lung cancer who might benefit from sequential tyrosine kinase inhibitor treatment.

BACKGROUND: Survival data support the use of first-line osimertinib as the standard of care for epidermal growth factor receptor (EGFR)-positive non-small cell lung cancer (NSCLC). However, it remains unclear whether upfront osimertinib is superior to sequential first- or second-generation tyrosine kinase inhibitors (TKIs) followed by osimertinib for all patients. It is impossible to predict which patients are at high risk of progression, and this constitutes a major limitation of the sequential TKI approach. PATIENTS AND METHODS: A total of 830 plasma samples from 228 patients with stage IV, EGFR-positive NSCLC who were treated with first-line TKIs were analysed by digital polymerase chain reaction (dPCR). RESULTS: The circulating tumour DNA (ctDNA) levels helped to identify patients with significantly improved survival rate, regardless of the treatment. Patients treated with first- or second-generation TKIs (N = 189) with EGFR mutations in plasma at a mutant allele frequency (MAF) <7% before treatment initiation (low-risk patients) or who were ctDNA negative after 3 or 6 months of treatment and with an MAF <7% at diagnosis (high responders) had two-thirds lower risk of death than patients in the opposite situation (adjusted hazard ratio [HR] = 0.38; 95% confidence interval [CI]: 0.23-0.64 and HR = 0.22; 95% CI: 0.12-0.42, respectively). The median overall survival (OS) for low-risk patients and high responders treated with first- or second-generation TKIs was 34.2 months and not reached, respectively, regardless of second-line treatment. There were no significant difference in OS between low-risk or high-responder patients treated upfront with osimertinib (N = 39) and those treated under a sequential approach with osimertinib (N = 60). Median OS was not reached in both cases. CONCLUSIONS: Pre-treatment ctDNA levels identify low-risk patients, who may benefit from sequential TKI treatment. Information regarding EGFR mutation clearance can help to improve patient selection.