Related Experiment Videos
Improved detection of prostate cancer using classification and regression tree analysis
Mark Garzotto1, Tomasz M Beer, R Guy Hudson
1Urology Section, Division of Urology, Portland Veterans Administration Medical Center, 3710 SW US Veterans Hospital Rd, Portland, OR 97239,USA. garzotto@ohsu.edu
Summary
This study developed a decision tree using Classification and Regression Trees (CART) analysis to reduce unnecessary prostate cancer biopsies. The CART model effectively identifies patients needing further work-up while maintaining high sensitivity for cancer detection.
Area of Science:
- Urology
- Oncology
- Medical Informatics
Background:
- Prostate cancer diagnosis often relies on prostate-specific antigen (PSA) levels and digital rectal examination (DRE), leading to a high rate of unnecessary biopsies.
- Developing more accurate diagnostic tools is crucial for efficient patient management and resource allocation.
Purpose of the Study:
- To construct a decision tree model for guiding prostate biopsy decisions in patients with suspected prostate cancer.
- To utilize Classification and Regression Tree (CART) analysis for improved diagnostic accuracy.
Main Methods:
- Collected data from 1,433 men with PSA levels ≤ 10 ng/mL undergoing prostate biopsy.
- Employed CART analysis, initially with PSA and DRE, then incorporating demographic, laboratory, and ultrasound data (e.g., hypoechoic lesions, PSA density [PSAD]).
- Validated the model using 20% of the data, assessing specificity, sensitivity, and accuracy via Receiver Operator Characteristic (ROC) curve analysis.
Main Results:
- CART identified a PSA cutoff > 1.55 ng/mL for further evaluation, irrespective of DRE results.
- Specific subgroups at risk for positive biopsy were defined based on PSAD, hypoechoic lesions, age, and prostate volume.
- The CART model demonstrated high sensitivity (96.6%) and comparable accuracy to logistic regression (AUC = 0.74 vs. 0.72), though with lower specificity (31.3%).
Conclusions:
- CART analysis offers a significant reduction in unnecessary prostate biopsies compared to standard PSA/DRE screening.
- The developed decision tree retains high sensitivity for detecting prostate cancer, particularly for clinically significant cancers.
- This approach enhances the efficiency of prostate cancer diagnosis by optimizing biopsy decisions.