PURPOSE: Population-based cancer registries (PBCRs) collect data on all new cancer diagnoses in a defined population. Data are sourced from pathology reports, and the PBCRs rely on manual and rule-based solutions. This study presents a state-of-the-art natural language processing (NLP) pipeline, built by fine-tuning pretrained language models (LMs). The pipeline is deployed at the British Columbia Cancer Registry (BCCR) to detect reportable tumors from a population-based feed of electronic pathology. METHODS: We fine-tune two publicly available LMs, GatorTron and BlueBERT, which are pretrained on clinical text. Fine-tuning is done using BCCR's pathology reports. For the final decision making, we combine both models' output using an OR approach. The fine-tuning data set consisted of 40,000 reports from the diagnosis year of 2021, and the test data sets consisted of 10,000 reports from the diagnosis year 2021, 20,000 reports from diagnosis year 2022, and 400 reports from diagnosis year 2023. RESULTS: The retrospective evaluation of our proposed approach showed boosted reportable accuracy, maintaining the true reportable threshold of 98%. CONCLUSION: Disadvantages of rule-based NLP in cancer surveillance include manual effort in rule design and sensitivity to language change. Deep learning approaches demonstrate superior performance in classification. PBCRs distinguish reportability status of incoming electronic cancer pathology reports. Deep learning methods provide significant advantages over rule-based NLP.

Background: The growing use of primary human papillomavirus (HPV) cervical cancer screening requires determining appropriate screening intervals to avoid overtreatment of transient disease. This study examined the long-term risk of cervical precancer after HPV screening to inform screening interval recommendations. Methods: This longitudinal cohort study (British Columbia, Canada, 2008 to 2022) recruited women and individuals with a cervix who received 1 to 2 negative HPV screens (HPV1 cohort, N = 5,546; HPV2 cohort, N = 6,624) during a randomized trial and women and individuals with a cervix with 1 to 2 normal cytology results (BCS1 cohort, N = 782,297; BCS2 cohort, N = 673,778) extracted from the provincial screening registry. All participants were followed through the registry for 14 years. Long-term risk of cervical precancer or worse [cervical intraepithelial neoplasia grade 2 or worse (CIN2+)] was compared between HPV and cytology cohorts. Results: Cumulative risks of CIN2+ were 3.2/1,000 [95% confidence interval (CI), 1.6–4.7] in HPV1 and 2.7/1,000 (95% CI, 1.2–4.2) in HPV2 after 8 years. This was comparable with the risk in the cytology cohorts after 3 years [BCS1: 3.3/1,000 (95% CI, 3.1–3.4); BCS2: 2.5/1,000 (95% CI, 2.4–2.6)]. The cumulative risk of CIN2+ after 10 years was low in the HPV cohorts [HPV1: 4.7/1,000 (95% CI, 2.6–6.7); HPV2: 3.9 (95% CI, 1.1–6.6)]. Conclusions: Risk of CIN2+ 8 years after a negative screen in the HPV cohorts was comparable with risk after 3 years in the cytology cohorts (the benchmark for acceptable risk). Impact: These findings suggest that primary HPV screening intervals could be extended beyond the current 5-year recommendation, potentially reducing barriers to screening.

PURPOSE: Rapid advancements in natural language processing have led to the development of sophisticated language models. Inspired by their success, these models are now used in health care for tasks such as clinical documentation and medical record classification. However, language models are prone to errors, which can have serious consequences in critical domains such as health care, ensuring that their reliability is essential to maintain patient safety and data integrity. METHODS: To address this, we propose an innovative auditing process based on principles from clinical trial design. Our approach involves subject matter experts (SMEs) manually reviewing pathology reports without previous knowledge of the model's classification. This single-blind setup minimizes bias and allows us to apply statistical rigor to assess model performance. RESULTS: Deployed at the British Columbia Cancer Registry, our audit process effectively identified the core issues in the operational models. Early interventions addressed these issues, maintaining data integrity and patient care standards. CONCLUSION: The audit provides real-world performance metrics and underscores the importance of human-in-the-loop machine learning. Even advanced models require SME oversight to ensure accuracy and reliability. To our knowledge, we have developed the first continuous audit process for language models in health care, modeled after clinical trial principles. This methodology ensures that audits are statistically sound and operationally feasible, setting a new standard for evaluating language models in critical applications.

Objective: Healthcare data fragmentation presents a major challenge for linking patient data, necessitating robust record linkage to integrate patient records from diverse sources. This study investigates the feasibility of leveraging language models for automated patient record linkage, focusing on two key tasks: blocking and matching. Materials and Methods: We utilized real-world healthcare data from the Missouri Cancer Registry and Research Center, linking patient records from two independent sources using probabilistic linkage as a baseline. A transformer-based model, RoBERTa, was fine-tuned for blocking using sentence embeddings. For matching, several language models were experimented under fine-tuned and zero-shot settings, assessing their performance against ground truth labels. Results: The fine-tuned blocking model achieved a 92% reduction in the number of candidate pairs while maintaining near-perfect recall. In the matching task, fine-tuned Mistral-7B achieved the best performance with only 6 incorrect predictions. Among zero-shot models, Mistral-Small-24B performed best, with a total of 55 incorrect predictions. Discussion: Fine-tuned language models achieved strong performance in patient record blocking and matching with minimal errors. However, they remain less accurate and efficient than a hybrid rule-based and probabilistic approach for blocking. Additionally, reasoning models like DeepSeek-R1 are impractical for large-scale record linkage due to high computational costs. Conclusion: This study highlights the potential of language models for automating patient record linkage, offering improved efficiency by eliminating the manual efforts required to perform patient record linkage. Overall, language models offer a scalable solution that can enhance data integration, reduce manual effort, and support disease surveillance and research.

Supplementary Table S1. Crude cumulative risk of CIN2+ and CIN3+ at intervals over follow-up among WIC with one or two negative HPV screen(s)

Background:The growing use of primary human papillomavirus (HPV) cervical cancer screening requires determining appropriate screening intervals to avoid overtreatment of transient disease. This study examined the long-term risk of cervical precancer after HPV screening to inform screening interval recommendations.Methods:This longitudinal cohort study (British Columbia, Canada, 2008 to 2022) recruited women and individuals with a cervix who received 1 to 2 negative HPV screens (HPV1 cohort, N = 5,546; HPV2 cohort, N = 6,624) during a randomized trial and women and individuals with a cervix with 1 to 2 normal cytology results (BCS1 cohort, N = 782,297; BCS2 cohort, N = 673,778) extracted from the provincial screening registry. All participants were followed through the registry for 14 years. Long-term risk of cervical precancer or worse [cervical intraepithelial neoplasia grade 2 or worse (CIN2+)] was compared between HPV and cytology cohorts.Results:Cumulative risks of CIN2+ were 3.2/1,000 [95% confidence interval (CI), 1.6–4.7] in HPV1 and 2.7/1,000 (95% CI, 1.2–4.2) in HPV2 after 8 years. This was comparable with the risk in the cytology cohorts after 3 years [BCS1: 3.3/1,000 (95% CI, 3.1–3.4); BCS2: 2.5/1,000 (95% CI, 2.4–2.6)]. The cumulative risk of CIN2+ after 10 years was low in the HPV cohorts [HPV1: 4.7/1,000 (95% CI, 2.6–6.7); HPV2: 3.9 (95% CI, 1.1–6.6)].Conclusions:Risk of CIN2+ 8 years after a negative screen in the HPV cohorts was comparable with risk after 3 years in the cytology cohorts (the benchmark for acceptable risk).Impact:These findings suggest that primary HPV screening intervals could be extended beyond the current 5-year recommendation, potentially reducing barriers to screening.