From the Journals

Automated bone scan index predicts survival in metastatic castration-resistant prostate cancer patients


Key clinical point: For patients with metastatic castration-resistant prostate cancer, an automated bone scan index (BSI) system predicted overall survival even after adjustment for other known prognostic factors.

Major finding: In multivariate analysis, higher automated BSI values remained independently associated with overall survival (hazard ratio, 1.06; 95% CI, 1.01-1.11; P = .03).

Study details: Preplanned secondary analysis of automated BSI in 721 men with bone metastatic, chemotherapy-naive castration-resistant prostate cancer who were enrolled in 10TASQ10, a randomized phase 3 clinical trial of tasquinimod.

Disclosures: The study was funded by EXINI Diagnostics AB, a subsidiary of Progenics Pharmaceuticals. The researchers reported disclosures related to EXINI Diagnostics AB and Active Biotech.

Source: Armstrong AJ et al. JAMA Oncol. 2018 May 17. doi: 10.1001/jamaoncol.2018.1093.

View on the News

Findings validate utility of automated BSI

This study prospectively validates the potential usefulness and establishes the prognostic value of an automated system for evaluating bone metastases, according to Fred Saad, MD.

The automated bone scan index (aBSI) was predictive of overall survival in univariate analysis, and remained significant in multivariate analysis, Dr. Saad noted in an editorial accompanying the study.

Despite some limitations, the system provides an encouraging framework to build on, he noted.

“With ongoing and future work in validating aBSI, this technology could eventually be introduced in clinical practice and research settings for improved stratification,” he wrote in the editorial.

Now, researchers are integrating Prostate Cancer Working Group progression criteria into the aBSI framework, which could be useful for evaluating response and resistance to treatment, if validated.

“This is extremely important for anyone who realizes the time and effort required for comparing bone scans to determine whether or not patients are progressing on clinical trials,” Dr. Saad wrote.

Dr. Saad is the University of Montreal Endowed Chair in Prostate Cancer, Surgery, University of Montreal Hospital Center. These comments are derived from his editorial in JAMA Oncology . Dr. Saad had no reported conflict of interest related to the editorial.


FROM JAMA Oncology

A system that automatically evaluates bone scans produces a measure that is prognostic for overall survival (OS) in patients with metastatic prostate cancer, investigators are reporting.

The automated bone scan index (BSI) produced by the artificial intelligence–based system provides prognostic discrimination beyond other established risk markers, investigators reported in JAMA Oncology.

The automated BSI also provides significantly greater discriminative ability compared with the current standard, which is manual assessment by counting metastatic lesion numbers, according to Andrew J. Armstrong, MD, of Duke University, Durham, N.C. and his coauthors.

“Incorporating the aBSI into clinical practice to supplement nuclear medicine reports may permit a more objective analysis of bone scan changes over time and their clinical relevance to patient care,” Dr. Armstrong and colleagues said.

The BSI represents tumor burden as a percentage of total skeletal weight. The system used in this study automates BSI methods by using artificial neural networks to detect metastatic hot spots and classify them as malignant or benign.

The automated system was evaluated in a preplanned secondary analysis of a randomized clinical trial of imiquimod in men with chemotherapy-naive metastatic castration-resistant prostate cancer. Evaluable bone scans were available for 721 out of 1,245 total enrollees from 241 sites in 37 countries.

In the secondary analysis, investigators found that baseline aBSI was significantly associated with OS. Risk of death increased by 20% for every doubling of the aBSI score (hazard ratio, 1.20; 95% confidence interval, 1.14-1.26; P less than .001).

Next Article: