Objective: This study explores a semi-supervised learning (SSL), pseudo-labeled strategy using diverse datasets such as head and neck cancer (HNCa) to enhance lung cancer (LCa) survival outcome predictions, analyzing handcrafted and deep radiomic features (HRF/DRF) from PET/CT scans with hybrid machine learning systems (HMLSs). Methods: We collected 199 LCa patients with both PET and CT images, obtained from TCIA and our local database, alongside 408 HNCa PET/CT images from TCIA. We extracted 215 HRFs and 1024 DRFs by PySERA and a 3D autoencoder, respectively, within the ViSERA 1.0.0 software, from segmented primary tumors. The supervised strategy (SL) employed an HMLS–PCA connected with six classifiers on both HRFs and DRFs. The SSL strategy expanded the datasets by adding 408 pseudo-labeled HNCa cases (labeled by the Random Forest algorithm) to 199 LCa cases, using the same HMLS techniques. Furthermore, principal component analysis (PCA) linked with four survival prediction algorithms were utilized in the survival hazard ratio analysis. Results: The SSL strategy outperformed the SL method (p << 0.001), achieving an average accuracy of 0.85 ± 0.05 with DRFs from PET and PCA + Multi-Layer Perceptron (MLP), compared to 0.69 ± 0.06 for the SL strategy using DRFs from CT and PCA + Light Gradient Boosting (LGB). Additionally, PCA linked with Component-wise Gradient Boosting Survival Analysis on both HRFs and DRFs, as extracted from CT, had an average C-index of 0.80, with a log rank p-value << 0.001, confirmed by external testing. Conclusions: Shifting from HRFs and SL to DRFs and SSL strategies, particularly in contexts with limited data points, enabling CT or PET alone, can significantly achieve high predictive performance.

Background: Patients with cancer often have unmet psychosocial needs. Early detection of who requires referral to a counsellor or psychiatrist may improve their care. This work used natural language processing to predict which patients will see a counsellor or psychiatrist from a patient’s initial oncology consultation document. We believe this is the first use of artificial intelligence to predict psychiatric outcomes from non-psychiatric medical documents. Methods: This retrospective prognostic study used data from 47,625 patients at BC Cancer. We analyzed initial oncology consultation documents using traditional and neural language models to predict whether patients would see a counsellor or psychiatrist in the 12 months following their initial oncology consultation. Results: Here, we show our best models achieved a balanced accuracy (receiver-operating-characteristic area-under-curve) of 73.1% (0.824) for predicting seeing a psychiatrist, and 71.0% (0.784) for seeing a counsellor. Different words and phrases are important for predicting each outcome. Conclusion: These results suggest natural language processing can be used to predict psychosocial needs of patients with cancer from their initial oncology consultation document. Future research could extend this work to predict the psychosocial needs of medical patients in other settings.

Adjuvant durvalumab after chemoradiotherapy (CRT) is the standard of care for unresectable stage III non-small cell lung cancer (NSCLC). A post hoc exploratory analysis of PACIFIC revealed no OS benefit in the PD-L1 < 1% subgroup. This retrospective analysis assesses the real-world impact of durvalumab on OS according to PD-L1 tumor proportion score (TPS). Patients with stage III, unresectable NSCLC treated by CRT, with available PD-L1 TPS, from 1 March 2018 to 31 December 2020, at BC Cancer, British Columbia, Canada were included. Patients were divided into two groups, CRT + durvalumab and CRT alone. OS and PFS were analyzed in the PD-L1 ≥ 1% and <1% subgroups. A total of 134 patients were included in the CRT + durvalumab group and 117, in the CRT alone group. Median OS was 35.9 months in the CRT + durvalumab group and 27.4 months in the CRT alone group [HR 0.59 (95% CI 0.42–0.83), p = 0.003]. Durvalumab improved OS in the PD-L1 ≥ 1% [HR 0.53 (95% CI 0.34–0.81), p = 0.003, n = 175], but not in the <1% subgroup [HR 0.79 (95% CI 0.44–1.42), p = 0.4, n = 76]. This retrospective study demonstrates a statistically significant improvement in OS associated with durvalumab after CRT in PD-L1 ≥ 1%, but not PD-L1 < 1% NSCLC. Variables not accounted for may have biased the survival analysis. A prospective study would bring more insight.

Background: A fixed dose of 200 mg of pembrolizumab every 3 weeks (Q3W) is the standard of care for patients with stage IV non-small cell lung cancer (NSCLC) and PDL1 ≥50%. In April 2020, based on pharmacokinetic modeling without formal comparative studies, the FDA approved 400 mg every 6 weeks (Q6W). Pharmacokinetic studies also suggested comparable target engagement with weight-based and flat dosing for the respective schedules. The objective of this study was to determine if overall survival (OS) differs based on the Q3W vs. Q6W dosing schedule of pembrolizumab. Methods: BC Cancer patients with stage IV NSCLC and PDL1 ≥50% treated with pembrolizumab were retrospectively reviewed. Patients were treated with weight-based dosing, per institution standard, of pembrolizumab 2 mg/kg Q3W or 4 mg/kg Q6W. Patient demographics, treatment and outcome were recorded. Patients were assigned to Q3W or Q6W according to the schedule that was used for the majority of treatment (greater than 50%). Results: 718 patients with NSCLC and PDL1 ≥50% received first-line pembrolizumab between 2017 and2021, Q3W/Q6W dosing 677/41 patients. Baseline characteristics with respect to age, sex, smoking status, histology and performance status (PS) were similar between groups. In the multivariate model, including age, sex, PS and dosing schedule, the hazard ratio for death (HR) for OS Q3W vs. Q6W was 0.759 (p = 0.230). A 2:1 case-matched analysis for OS was performed, controlling for sex, age ± 5 years, PS and duration on pembrolizumab ± 2 months for Q3W vs. Q6W (n = 113) with a HR 0.834 (p = 0.500). Conclusions: There was no OS difference demonstrated with pembrolizumab dosing Q3W compared to Q6W in a multivariate analysis that included age, sex and PS. A case-matched analysis that controlled for these variables and for duration of treatment confirmed these findings. This study supports the use of Q6W pembrolizumab dosing, allowing for less frequent interactions with the medical system.