Index Quantifies Bone Disease in Prostate Cancer

NEW YORK—Researchers at Memorial Sloan-Kettering Cancer Center havedeveloped a method of quantifying bone involvement in patients with androgen-independentprostate cancer and have found that the resulting bone scan index (BSI)correlates with patient survival. In contrast, simply counting the numberof bone lesions present did not provide useful prognostic information.

“At present, because of the difficulty in measuring bone disease, prostatecancer patients with bone-only disease are generally excluded from prostatecancer trials,” Paul Sabbatini, MD, said in an interview with OncologyNews International. “Since you can’t measure their disease, you can’t tellwhether a patient is getting better or worse with treatment.”

Yet, he pointed out, bone-only disease is the most common prostate cancerpresentation, and these patients usually have earlier disease, “when youwould really like to test new strategies,” than those with measurable metastases,such as lymph node or soft tissue involvement. “Our ultimate goal,” hesaid, “is to develop a way to make bone disease measurable via serial bonescans that can be compared objectively over time.”

Calculating the BSI

To come up with a BSI, the radiologist visually estimates the percentinvolvement of each bone, based on a patient’s bone scans. “This is lesslaborious than you might think,” Dr. Sabbatini said. “Usually, you havea total of seven or eight lesions. You might note, for example, that 20%of the clavicle is involved. That percentage is then multiplied by theclavicle’s percentage of the total weight of the skeleton.”

These proportions are already known from a published “reference man”that shows the proportional weight of about 120 bones in the average humanskeleton. “When this is done for each bone lesion, you come up with a percentageof total bone involvement, termed the BSI,” he said. The higher the BSI,the greater the bone involvement (see Figure).

In a so-called superscan in which almost the entire skeleton shows tumorinvolvement, the BSI is 50% to 60%. Serial BSI measurements would not beuseful in these patients, Dr. Sabbatini said, and these patients typicallyalso have lymph node or liver disease that can be measured in therapeutictrials. (Even in a “superscan,” he noted, the BSI is not 100%, becauseof the allowance for bone marrow and for areas of bone where tumor doesnot go.)

In their study, presented at the American Society of Clinical Oncology1997 meeting, the investigators calculated baseline BSI in 191 prostatecancer patients who had progressive disease after primary hormonal therapy.These patients were enrolled in a randomized phase III study of oral liarozolevs predni-sone, and survival data were available.

The patients were stratified into three equal groups based on baselineBSI: less than 1.4% (63 patients), 1.4% to 5.1% (65 patients), and morethan 5.1% (63).

In a multivariate analysis controlling for other prognostic factorssuch as baseline PSA, age, LDH, and treatment arm, BSI was shown to beindependently predictive of survival. Among those with the lowest baselineBSI, median survival was 18.3 months vs 15.5 months and 8.1 months, respectively,in the two groups with higher baseline BSI (P = .0079).

Further studies are planned to define the prognostic significance ofserial BSI measurements in prostate cancer patients with bone lesions,Dr. Sabbatini said. “Although we did not have serial bone scans for thedataset used in the current study, we did have survival data, and we thoughtit was important first to be able to say definitively that BSI correlateswith survival, and we were able to do that. The next step is to look atserial scans.”