Is it possible to automatically assess pretreatment digital rectal examination documentation using natural language processing? A single-centre retrospective study.

OBJECTIVES: To develop and test a method for automatic assessment of a quality metric, provider-documented pretreatment digital rectal examination (DRE), using the outputs of a natural language processing (NLP) framework. SETTING: An electronic health records (EHR)-based prostate cancer data warehouse was used to identify patients and associated clinical notes from 1 January 2005 to 31 December 2017. Using a previously developed natural language processing pipeline, we classified DRE assessment as documented (currently or historically performed), deferred (or suggested as a future examination) and refused. PRIMARY AND SECONDARY OUTCOME MEASURES: We investigated the quality metric performance, documentation 6 months before treatment and identified patient and clinical factors associated with metric performance. RESULTS: The cohort included 7215 patients with prostate cancer and 426 227 unique clinical notes associated with pretreatment encounters. DREs of 5958 (82.6%) patients were documented and 1257 (17.4%) of patients did not have a DRE documented in the EHR. A total of 3742 (51.9%) patient DREs were documented within 6 months prior to treatment, meeting the quality metric. Patients with private insurance had a higher rate of DRE 6 months prior to starting treatment as compared with Medicaid-based or Medicare-based payors (77.3%vs69.5%, p=0.001). Patients undergoing chemotherapy, radiation therapy or surgery as the first line of treatment were more likely to have a documented DRE 6 months prior to treatment. CONCLUSION: EHRs contain valuable unstructured information and with NLP, it is feasible to accurately and efficiently identify quality metrics with current documentation clinician workflow.

Comparison of orthogonal NLP methods for clinical phenotyping and assessment of bone scan utilization among prostate cancer patients.

OBJECTIVE: Clinical care guidelines recommend that newly diagnosed prostate cancer patients at high risk for metastatic spread receive a bone scan prior to treatment and that low risk patients not receive it. The objective was to develop an automated pipeline to interrogate heterogeneous data to evaluate the use of bone scans using a two different Natural Language Processing (NLP) approaches. MATERIALS AND METHODS: Our cohort was divided into risk groups based on Electronic Health Records (EHR). Information on bone scan utilization was identified in both structured data and free text from clinical notes. Our pipeline annotated sentences with a combination of a rule-based method using the ConText algorithm (a generalization of NegEx) and a Convolutional Neural Network (CNN) method using word2vec to produce word embeddings. RESULTS: A total of 5500 patients and 369,764 notes were included in the study. A total of 39% of patients were high-risk and 73% of these received a bone scan; of the 18% low risk patients, 10% received one. The accuracy of CNN model outperformed the rule-based model one (F-measure = 0.918 and 0.897 respectively). We demonstrate a combination of both models could maximize precision or recall, based on the study question. CONCLUSION: Using structured data, we accurately classified patients' cancer risk group, identified bone scan documentation with two NLP methods, and evaluated guideline adherence. Our pipeline can be used to provide concrete feedback to clinicians and guide treatment decisions.

Positive family history of prostate cancer not associated with worse outcomes after radical prostatectomy.

OBJECTIVES: To determine the clinical outcomes in men with (FH) and without (NFH) a family history of prostate cancer after radical prostatectomy. METHODS: We performed a retrospective analysis of 557 men with localized prostate cancer treated by radical prostatectomy between 1989 and 2000. We defined a positive FH as having one or more first-degree relatives such as a father or brother with prostate cancer. The clinical and pathologic features, as well as biochemical disease-free survival, defined as an undetectable prostate-specific antigen level (less than 0.2 ng/mL), were compared between the FH and NFH groups. RESULTS: Compared with the NFH group, the FH men were younger at surgery (median 62 years versus 64 years, P = 0.01), had a lower median preoperative prostate-specific antigen level (7.2 ng/mL versus 7.8 ng/mL, P = 0.05), and were more likely to have only low-grade disease at the final pathologic evaluation (26.2% versus 17.8%, P = 0.05). At a median follow-up of 7.5 years (mean 7.6 +/- 2.9 years), 17% of the FH group had biochemical disease recurrence compared with 30% in the NFH group. The actuarial disease-free survival rate at 5 and 10 years for the two groups was 86% and 80% compared with 73% and 66%, respectively (P = 0.01). When controlled for pathologic variables in a multivariate analysis, FH was not an independent predictor of disease-free survival. CONCLUSIONS: The association of improved disease-free survival in the FH patients may have been driven by an earlier age at diagnosis and more favorable pathologic features.

Vascular invasion predicts recurrence after radical prostatectomy: stratification of risk based on pathologic variables.

OBJECTIVES: To determine whether vascular invasion (VI) is an independent predictor of prostate cancer recurrence and/or survival and to stratify risk of recurrence in patients with VI. METHODS: Vascular invasion status was documented in 620 radical prostatectomy specimens with an average of 7.5 years of follow-up. The relationship between VI and other clinical and pathologic features was tested. Vascular invasion as an independent predictor of recurrence was investigated by logistic regression analysis. Survival analyses and stratification of VI patients was developed with Kaplan-Meier survival curves. RESULTS: Vascular invasion was identified in 110 patients (18%) and correlated significantly (P <0.0001) with high Gleason grade, extracapsular extension (EPE), seminal vesicle invasion, increasing cancer volumes, positive margins, and elevated preoperative prostate-specific antigen (PSA) levels. Logistic regression analysis demonstrated that VI was a strong and independent predictor for disease recurrence, when considered with grade, EPE, seminal vesicle invasion, lymph node involvement, cancer volume, preoperative PSA levels, and positive margins. At 12 years after radical prostatectomy, patients with VI demonstrated significantly lower disease-specific survival (P = 0.0005). Among patients with VI, stratification of grade, EPE, and the number of VI foci identified three significantly different prognostic groups. CONCLUSIONS: In long-term follow-up, VI was a significant predictor of prostate cancer recurrence and death after radical prostatectomy. In patients with VI, high Gleason grade, EPE, and more than five foci of VI are associated with poor prognosis.

Time trends in pathologic features of radical prostatectomy--impact of family history.

We investigated whether the clinical or pathological features of patients with a family history of prostate cancer treated by radical prostatectomy differ from patients without a family history. A retrospective analysis of patients treated by radical prostatectomy between 1989 through 2000 was performed. The clinical and pathologic features of patients with a family history (defined as at least one first-degree relative with prostate cancer, N = 103) were compared with those with no family history (N = 456). In addition, the patients were stratified into two groups, those treated from 1989 through 1992 and those treated after 1992. In the entire cohort from 1989 through 2000, patients with a family history had a greater proportion of well-differentiated tumors than the NFH group (26.2% vs. 17.8%; P = 0.05). From 1989 to 1992 there was no statistical difference between patients with a family history (FH) and those without a family history (NFH) with respect to age, prostate specific antigen (PSA), PSA density, clinical or pathologic stage, Gleason grade, or total tumor volume. However, after 1992 the FH group tended to be younger than the NFH group (61.1 vs. 63.4; P = 0.02) and have a lower PSA (6.8 vs. 7.9; P = 0.01) at the time of diagnosis. We believe these differences are predominantly driven by more aggressive screening in patients with a family history of prostate cancer rather than any true genetic differences.