Diagnostic efficacy of free to total ratio of prostate-specific antigen and prostate-specific antigen velocity, singly and in combination, in detecting prostate cancer in patients with total serum prostate-specific antigen between 4 and 10 ng/ml

Advancing Traditional Prostate-specific Antigen Kinetics in the Detection of Prostate Cancer: A Machine Learning Model

BACKGROUND

Prostate-specific antigen (PSA) kinetics, defined as the change in PSA over time, may be of use as a predictor of prostate cancer. PSA kinetics can be assessed as the PSA velocity, which is traditionally evaluated dichotomously and classified as abnormal if greater than either 0.35 or 0.75 ng/ml/yr. Machine learning models may provide additional benefit in assessing risk using PSA kinetics instead of PSA velocity.

OBJECTIVE

To improve the utility of PSA kinetics by constructing a generalizable, universal machine learning model.

DESIGN, SETTING, AND PARTICIPANTS

Data were obtained from the PLCO and PCPT trials and from a contemporary Australian cohort. PSA data were interpolated using a modified Gaussian process. A machine learning model based on a two-headed approach was designed, in which the multivariable input was fed into a one-dimensional ResNet18 model.

OUTCOME MEASURES AND STATISTICAL ANALYSIS

The model performance was assessed compared to PSA levels and PSA velocity in terms of area under the receiver operator characteristic curve (AUC).

RESULTS AND LIMITATIONS

A total of 10719 patients were included in the analysis. In tests on a validation set of the complete database to diagnose grade group ≥2, the AUC was 0.886 (95% confidence interval [CI] 0.870-0.902) for the machine learning model, compared to 0.807 (95% CI 0.796-0.819) for PSA and 0.627 (95% CI 0.607-0.648) for PSA velocity.

CONCLUSIONS

Machine learning models can be used to augment the diagnostic utility of PSA kinetics in the diagnosis of prostate cancer. We demonstrated significant improvements in accuracy compared to the traditional approaches of PSA velocity and PSA thresholds.

PATIENT SUMMARY

Prostate cancer diagnosis is limited by the diagnostic accuracy of the prostate-specific antigen (PSA) blood test. Advances in techniques such as machine learning algorithms can greatly improve the diagnostic accuracy of prostate cancer screening without additional costs or tests.

Multiplexed Prostate Cancer Companion Diagnostic Devices

Prostate cancer (PCa) remains one of the most prominent forms of cancer for men. Since the early 1990s, Prostate-Specific Antigen (PSA) has been a commonly recognized PCa-associated protein biomarker. However, PSA testing has been shown to lack in specificity and sensitivity when needed to diagnose, monitor and/or treat PCa patients successfully. One enhancement could include the simultaneous detection of multiple PCa-associated protein biomarkers alongside PSA, also known as multiplexing. If conventional methods such as the enzyme-linked immunosorbent assay (ELISA) are used, multiplexed detection of such protein biomarkers can result in an increase in the required sample volume, in the complexity of the analytical procedures, and in adding to the cost. Using companion diagnostic devices such as biosensors, which can be portable and cost-effective with multiplexing capacities, may address these limitations. This review explores recent research for multiplexed PCa protein biomarker detection using optical and electrochemical biosensor platforms. Some of the novel and potential serum-based PCa protein biomarkers will be discussed in this review. In addition, this review discusses the importance of converting research protocols into multiplex point-of-care testing (xPOCT) devices to be used in near-patient settings, providing a more personalized approach to PCa patients’ diagnostic, surveillance and treatment management.

Value of free/total prostate-specific antigen (f/t PSA) ratios for prostate cancer detection in patients with total serum prostate-specific antigen between 4 and 10 ng/mL: A meta-analysis

BACKGROUND

Prostate carcinoma is a common disease that occurs in men over 50 years old. Many studies have explored the effect of free/total prostate-specific antigen (f/t PSA) ratio in monitoring prostate cancer. We conducted a meta-analysis to identify the accuracy of the f/t PSA ratio in the diagnosis of prostate cancer in patients who have PSA levels of 4 to 10 ng/mL.

METHODS

Databases searched included PubMed and OVID databases, from inception to March 2017, after a systematical review, sensitivity, specificity, and other measures of accuracy of the f/t PSA ratio in the diagnosis of prostate cancer were pooled. We used summary receiver operating characteristic curves to summarize overall test performance.

RESULTS

Fifteen case-control studies from 14 articles were identified. The results indicated that the sensitivity of the f/t PSA ratio in the diagnosis of prostate cancer ranged from 0.5 to 0.94 (pooled sensitivity 0.70, 95% CI: 0.67-0.72), whereas its specificity ranged from 0.31 to 0.93 (pooled specificity 0.55, 95% CI: 0.57-0.60). The positive likelihood ratio was 1.85 (95% CI: 1.56-2.20), negative likelihood ratio was 0.42 (95% CI: 0.34-0.53), and diagnostic odds ratio was 4.81 (9.53% CI: 3.33-6.94).

CONCLUSIONS

The f/t PSA ratio determination has a low sensitivity and specificity for the diagnosis of prostate cancer; it would not be useful for the diagnosis of prostate cancer by itself. The results of f/t PSA ratio measurements should refer to the clinical manifestations and the results of conventional tests such as biopsies.

[Interpretation of laboratory results of elderly subjects: effect of age or aging]

The clinical pertinence of laboratory tests is conditioned by the use of well-documented databases on biological variability. Knowledge of these variations in the pediatric and adolescent population is well established. On the other hand, the situation is not so obvious for the elderly population. This population is generally not homogeneous and well-defined physiological variations are scarce. The observed variations generally reflect infraclinical or chronic illnesses. Interindividual variation is generally higher in the elderly population than in younger subjects. Thus, it would be more judicious to classify older persons as a function of their functional performance. This raises the issue of the proper definition of "good health" during the aging process. The criteria usually used for young adults should be revisited. The relationship with age is in fact variable from one assay to another: sometimes low, non significant or on contrary extremely wide. Well-adapted reference intervals are a prerequisite to proper interpretation of laboratory data in the elderly population. In some instances however, reference limits may be unsuitable, e.g. when the biological variability is too large, reference limits used for younger adults should be applied, while in other cases, traditional reference limits must be applied for decision-making limits.