Background

Active surveillance followed by selective treatment for men who have evidence of disease progression may be an option for select patients with early-stage prostate cancer. In this article, the authors report their experience in a contemporary cohort of men with prostate cancer who were managed with active surveillance.

Methods

All men who were managed initially with active surveillance were identified through the authors' institutional database. Selection criteria for active surveillance included: prostate-specific antigen (PSA)<10 ng/mL, biopsy Gleason sum 0.75 ng/mL per year), was a secondary outcome. Chi-square and log-rank tests were used to compare groups. The association between clinical characteristics and receipt of active treatment was analyzed by using Cox proportional hazards regression.

Results

Three hundred twenty-one men (mean age [+/-standard deviation]: 63.4+/-8.5 years) selected active surveillance as their initial management. The overall median follow-up was 3.6 years (range, 1-17 years). The initial mean PSA level was 6.5+/-3.9 ng/mL. One hundred twenty men (37%) met at least 1 criterion for progression. Overall, 38% of men had higher grade on repeat biopsy, and 26% of men had a PSA velocity>0.75 ng/mL per year. Seventy-eight men (24%) received secondary treatment at a median 3 years (range, 1-17 years) after diagnosis. Approximately 13% of patients with no disease progression elected to obtain treatment. PSA density at diagnosis and rise in Gleason score on repeat biopsy were associated significantly with receipt of secondary treatment. The disease-specific survival rate was 100%.

Conclusions

Selected individuals with early-stage prostate cancer may be candidates for active surveillance. Specific criteria can be and need to be developed to select the most appropriate individuals for this form of management and to monitor disease progression. A small attrition rate can be expected because of men who are unable or unwilling to tolerate surveillance.

Purpose

Cancer pathology findings are critical for many aspects of care but are often locked away as unstructured free text. Our objective was to develop a natural language processing (NLP) system to extract prostate pathology details from postoperative pathology reports and a parallel structured data entry process for use by urologists during routine documentation care and compare accuracy when compared with manual abstraction and concordance between NLP and clinician-entered approaches.

Materials and methods

From February 2016, clinicians used note templates with custom structured data elements (SDEs) during routine clinical care for men with prostate cancer. We also developed an NLP algorithm to parse radical prostatectomy pathology reports and extract structured data. We compared accuracy of clinician-entered SDEs and NLP-parsed data to manual abstraction as a gold standard and compared concordance (Cohen's κ) between approaches assuming no gold standard.

Results

There were 523 patients with NLP-extracted data, 319 with SDE data, and 555 with manually abstracted data. For Gleason scores, NLP and clinician SDE accuracy was 95.6% and 95.8%, respectively, compared with manual abstraction, with concordance of 0.93 (95% CI, 0.89 to 0.98). For margin status, extracapsular extension, and seminal vesicle invasion, stage, and lymph node status, NLP accuracy was 94.8% to 100%, SDE accuracy was 87.7% to 100%, and concordance between NLP and SDE ranged from 0.92 to 1.0.

Conclusion

We show that a real-world deployment of an NLP algorithm to extract pathology data and structured data entry by clinicians during routine clinical care in a busy clinical practice can generate accurate data when compared with manual abstraction for some, but not all, components of a prostate pathology report.