Review of Palliative ²²³Ra in Metastatic Castration-Resistant Prostate Cancer: Experience at West Virginia University Cancer Center

DISCUSSION

The science supporting the development and clinical use of ²²³Ra is compelling. An α-emitter, ²²³Ra dichloride (²²³Ra) is bone-seeking and complexes with hydroxyapatite, releasing short-range (<100 μm), high-energy particles that exert a selective antitumor effect (20). Areas of active bone remodeling and bone vascular supply are the main target sites for ²²³Ra localization (21). In mouse models, ²²³Ra inhibits tumor-induced pathologic bone formation in the tumor microenvironment adjacent to tumor foci (22).

Preclinical work with mice confirmed that ²²³Ra selectively accumulates in the bone, with only minute amounts of daughter radionuclides released from skeletal sites of ²²³Ra decay (23). Moreover, dosimetry studies showed that short-range α-radiation from ²²³Ra substantially spares bone marrow, with relatively less toxicity than is caused by β-emitting ⁸⁹Sr (23).

A phase 1 clinical trial demonstrated that ²²³Ra preferentially targets bone metastases rather than diffusely targeting healthy bone tissue (24). A randomized, multicenter, placebo-controlled phase 2 trial assigned patients receiving external-beam radiation therapy for pain control to receive either 4 ²²³Ra injections (50 kBq kg⁻¹) at 4-wk intervals or placebo on the same schedule. The group receiving ²²³Ra demonstrated a significant decline in alkaline phosphatase levels, a delay in the time to prostate-specific antigen progression, and a trend toward reduced skeleton-related events and improved OS. The safety profile was acceptable, hematologic toxicity did not significantly differ between the 2 groups, and no patient discontinued ²²³Ra because of treatment-related toxicity (25).

Parker et al. reported the pivotal phase III study of ²²³Ra, involving 614 ²²³Ra-treated patients versus 307 in a placebo group, treated at 136 centers in 19 countries (3). At the time of publication, 532 of 921 patients (58%) had received all 6 injections of ²²³Ra. The median OS was 14.9 mo in the ²²³Ra group and 11.3 mo in the placebo group. ²²³Ra was associated with a 30% reduction in the risk of death versus placebo. The effect of ²²³Ra on OS was consistent across all subgroups, and ²²³Ra was not associated with significantly more grade 3 or 4 toxic effects than placebo.

However, subsequent to ²²³Ra approval and widespread use in the United States, multiple reports raised caveats concerning patient selection for ²²³Ra, finding shorter OS and more adverse events than reported by ALSYMPCA. Most of these have been smaller cohorts in single institutions, and there have also been a few multicenter experiences. Stolten et al. and Vogelzang et al. both looked at outcomes of patients receiving the full 6 cycles versus those who could not complete treatment. The cohorts had 55–184 patients, and OS favored those who completed all 6 cycles of therapy (16 mo vs. 4 mo in the Stolten et al. cohort, and 17 mo in the Vogelzang cohort) (4,5).

Etchebehere et al. reported 41 patients with metastatic prostate cancer treated at M.D. Anderson Cancer Center who had undergone fluoride PET/CT before ²²³Ra (26). They defined bone marrow failure as World Health Organization grade 3 hematologic toxicity with no recovery after 6 wk, or recorded death due to bone marrow failure. Bone marrow failure correlated with tumor burden on imaging, which, by multivariable analysis, was the only independent predictor for bone marrow failure. Another report by Etchebehere et al. on 110 ²²³Ra-treated patients noted that improved alkaline phosphatase levels and pain scores were associated with improved outcomes (6).

Additionally, there have also been multiple reports showing better outcomes when using ²²³Ra earlier in the course of disease. Hague et al., Saad et al., Baldari et al., Wong et al., and Sartor et al. all found this to be the case within their respective studies (8,12,27–29). Specifically, Saad et al. had the largest cohort of these and reported on the early-access program after the ALSYMPCA study and before regulatory approval of ²²³Ra in 2013. In total, 839 patients were enrolled from 113 sites in 14 countries; 696 patients received one or more doses of ²²³Ra; 403 (58%) had all 6 planned injections (28). Sartor et al. reported the U.S. experience from the expanded-access phase II open-label, single-arm multicenter trial. Those with more advanced disease were less likely to benefit from the isotope. The median OS in those receiving 5–6 injections was not reached, whereas the median survival in those with fewer than 5 injections was 7.5 mo (12).

In July 2018, the safety committee of the European Medicines Agency—the Pharmacovigilance Risk Assessment Committee—recommended restricting the use of ²²³Ra-dichloride to patients who had only 2 previous treatments for metastatic prostate cancer or cannot receive other treatments (13,14). The committee recommended that ²²³Ra not be used with abiraterone plus corticosteroid, with other systemic cancer therapies except for hormonal therapy, and in asymptomatic patients or those with few osteoblastic bone metastases. The committee further suggested that patients be carefully assessed for fracture risk and that use of bisphosphonates or RANKL inhibitors be considered to increase bone strength before ²²³Ra. Although O’Sullivan et al. made a case against these restrictive recommendations, the European Medicines Agency countered that the recommendations were not only because of detrimental effects on bone fractures when used with abiraterone but also because of risk of lymph node or visceral progression during ²²³Ra bone-focused treatment (30,31).

Of interest, one study has looked at possible predictive measures of survival in patients receiving ²²³Ra. Dittman et al. used quantitative bone SPECT/CT before ²²³Ra in 60 patients with mCRPC and found that isotope uptake in the central skeleton varied from 11% to 56% of the injected dose. After ²²³Ra, median OS in all 60 patients was 15.2 mo. When skeletal uptake was 26% or more, median OS was 7.3 mo; when skeletal uptake was less than 26%, median OS was 30.8 mo (17). The authors concluded that initial quantitative SPECT/CT bone scanning could predict subsequent patient survival after treatment with ²²³Ra. This finding may be of future interest once more data become available.

After reviewing our own patient experience, we identified certain factors likely contributing to our patients’ shorter-than-expected median OS. Practically all androgen deprivation, radiation, and cytotoxic therapy was given before referral to WVU for ²²³Ra. Although most referred patients initially seemed eligible for ²²³Ra on the basis of Eastern Cooperative Oncology Group performance status and the usual laboratory parameters, many appeared to receive only minor benefit from ²²³Ra as administered. When compared with the original ALSYMPCA prostate cancer patients, more patients at WVU received bisphosphonates and docetaxel, had a performance status of more than 2, took more opiates for cancer pain, had a greater bone metastasis burden, and had lower hemoglobin, albumin, alkaline phosphatase, and prostate-specific antigen levels (Tables 1–3).

Potential limitations of our study are that it was observational and retrospective, and our Appalachian referral-patient results may be hard to extrapolate to other practice settings. The ALSYMPCA trial was multicenter, multinational, and prospective, whereas our WVU data were from a single referral center. The number of ²²³Ra-treated patients studied in the 2 trials differed by over 25-fold, as we had a relatively small number of patients in our dataset (the last patient reported here started ²²³Ra in October 2018). Also, the data-collection time frames of the 2 studies differed by more than 5 y. Imaging in our patients included 12 nuclear bone scans (5 of which also included SPECT/CT), 8 body CT scans, and 4 PET/CT scans. Thus, only 7 patients had a nuclear bone scan, and use of higher-resolution imaging in the other 17 might have increased our detection of the number of bone lesions versus earlier studies using only conventional nuclear bone scans (17). Despite these issues, our report adds to the growing amount of real-world data concerning use of ²²³Ra, which will hopefully guide the future clinical application of this promising isotope.