Assessing Therapeutic Response to Radium-223 with an Automated Bone Scan Index among Metastatic Castration-Resistant Prostate Cancer Patients: Data from Patients in the J-RAP-BSI Trial

Simple Summary This study was a retrospective investigation of a Japanese cohort of 205 metastatic castration-resistant prostate cancer (mCRPC) patients who received Ra-223 in 14 hospitals between July 2016 and August 2020 and for whom bone scintigraphy before and after the radium-223 treatment was available. Following treatment, alkaline phosphatase (ALP) decline (%ALP < 0%) was noted in 72.2% (148/205), automated bone scan index (aBSI) decline (%aBSI < 0%) in 52.7% (108/205), and PSA decline (%PSA < 0%) in 27.8% (57/205). Furthermore, a reduction in both aBSI and ALP was seen in 87 (42.4%), a reduction in only ALP was seen in 61 (29.8%), a reduction in only aBSI was seen in 21 (10.2%), and in both aBSI and ALP increasing/stable (≥0%) was seen in 36 (17.6%) patients. Multiparametric analysis showed changes in PSA (HR 4.30, 95% CI 2.32–8.77, p < 0.0001), aBSI (HR 2.22, 95%CI 1.43–3.59, p = 0.0003), and ALP (HR 2.06, 95%CI 1.35–3.14, p = 0.0008) as significant prognostic factors for OS. For mCRPC patients treated with Ra-223, aBSI change is useful as an imaging biomarker for treatment response assessment and survival prediction. Abstract To evaluate the usefulness of change in the automated bone scan index (aBSI) value derived from bone scintigraphy findings as an imaging biomarker for the assessment of treatment response and survival prediction in metastatic castration-resistant prostate cancer (mCRPC) patients treated with Ra-223. This study was a retrospective investigation of a Japanese cohort of 205 mCRPC patients who received Ra-223 in 14 hospitals between July 2016 and August 2020 and for whom bone scintigraphy before and after radium-223 treatment was available. Correlations of aBSI change, with changes in the serum markers alkaline phosphatase (ALP) and prostate-specific antigen (PSA) were evaluated. Additionally, the association of those changes with overall survival (OS) was assessed using the Cox proportional-hazards model and Kaplan–Meier curve results. Of the 205 patients enrolled, 165 (80.5%) completed six cycles of Ra-223. Following treatment, ALP decline (%ALP < 0%) was noted in 72.2% (148/205), aBSI decline (%aBSI < 0%) in 52.7% (108/205), and PSA decline (%PSA < 0%) in 27.8% (57/205). Furthermore, a reduction in both aBSI and ALP was seen in 87 (42.4%), a reduction in only ALP was seen in 61 (29.8%), a reduction in only aBSI was seen in 21 (10.2%), and in both aBSI and ALP increasing/stable (≥0%) was seen in 36 (17.6%) patients. Multiparametric analysis showed changes in PSA [hazard ratio (HR) 4.30, 95% confidence interval (CI) 2.32–8.77, p < 0.0001], aBSI (HR 2.22, 95%CI 1.43–3.59, p = 0.0003), and ALP (HR 2.06, 95%CI 1.35–3.14, p = 0.0008) as significant prognostic factors for OS. For mCRPC patients treated with Ra-223, aBSI change is useful as an imaging biomarker for treatment response assessment and survival prediction.


Introduction
Radium-223 dichloride (Ra-223), a targeted alpha emitter, selectively binds to highbone turnover sites caused by bone metastasis [1]. Alpha particles with a very short range (<100 µm) primarily induce double-stranded DNA breaks, resulting in highly localized cytotoxic effects with lower levels of damage to surrounding tissues. The ALSYMPCA trial showed both the efficacy and safety of Ra-223 in metastatic castration-resistant prostate cancer (mCRPC) patients with symptomatic bone metastasis [2]. The results of the randomized phase 3 ALSYMPCA study, which compared patients with CRPC and symptomatic bone metastasis who received the best standard of care plus Ra-223 with those who received a placebo, showed that the former group had a median overall survival (OS) prolonged by 3.6 months [hazard ratio (HR) 0.70. 95% confidence interval (CI) 0.58-0.83, p < 0.001, median 14.9 vs. 11.3 months]. Additionally, the results indicated that Ra-223 treatment was well tolerated and resulted in a low incidence of grade 3 or 4 myelosuppression (Ra-223 vs. placebo: anemia 13% and 13%, neutropenia 2% and 1%, thrombocytopenia 7% and 2%).
For a proper evaluation of response to Ra-223 treatment, as well as stratification of responders and non-responders for subsequent therapy, the determination of radiographic response biomarkers is necessary. While alkaline phosphatase (ALP) and prostate specific antigen (PSA) levels are typically monitored [3], response biomarkers that provide better accuracy for choosing the most appropriate course of treatment would be helpful. Bone scintigraphy findings of mCRPC patients with bone metastasis, in whom bone uptake is proportional to bone remodeling activity, are fundamental for monitoring treatment in mCRPC cases and recommended by the Prostate Cancer Clinical Trials Working Group (PCWG3) for estimating skeletal response to therapy [4]. Nevertheless, the response to Ra-223 shown by bone scintigraphy and the correlation with clinical outcome factors have yet to be thoroughly elucidated.
To quantify the extent of skeletal tumor burden shown by bone scintigraphy as a percentage of total skeletal weight, an automated bone scan index (aBSI) (%) has been developed [5]. The important information it provides makes it a potentially helpful tool for estimating the total quantitative skeletal metastatic burden in mCRPC patients. This automated methodology utilizes artificial intelligence and has a rapid processing time, while it has also been demonstrated to provide accurate and reproducible results [6]. In a prospective phase III study of 721 mCRPC patients treated at multiple institutions recently reported, aBSI was clinically validated as a prognostic biomarker [7]. Additionally, several prior evaluations of pretreatment aBSI as a prognostic imaging biomarker for mCRPC patients who received Ra-223 treatment showed a significant association of baseline values with OS [8][9][10][11][12][13][14]. On the other hand, data regarding response biomarkers to evaluate treatment response or non-response shown by aBSI in mCRPC patients treated with Ra-223 are scarce. Two studies demonstrated that change in aBSI was a useful finding for survival prediction [10,12], whereas two others found no significance related to aBSI change [9,11]; thus, its usefulness remains controversial.
The Japanese Ra-223 Therapy in Prostate Cancer using Bone Scan Index (J-RAP-BSI) multicenter trial was performed to examine the effectiveness of aBSI as an imaging biomarker for the evaluation of bone metastasis in CRPC patients in Japan receiving Ra-223 therapy [14]. The present multicenter retrospective study was conducted as a sub-analysis of results obtained in that trial to evaluate aBSI findings for assessment of response to such therapy as well as survival prediction in mCRPC patients.

Ethics
body weight) administered every 28 days was given to mCRPC patients without disease progression, findings showing unacceptable toxicity, declining performance status, or by the request of the patient. The present cohort consisted of males >18 years old, with bone metastasis, no visceral metastasis or lymph node ≥3 cm, hemoglobin >8.4 gm/dl, white blood cell count >1.5 × 10 9 , and platelet count >100 × 10 9 at the time of the initial Ra-233 injection. Androgen deprivation therapy was continued during Ra-233 therapy, while other medications, such as denosumab or bisphosphonates, were prescribed at the discretion of the attending physician. Previous systemic therapy, such as abiraterone, enzalutamide, docetaxel, or cabazitaxel, was permitted for inclusion, while concomitant treatment with abiraterone or chemotherapy was not. Patients were excluded if they had undergone chemotherapy within the previous four weeks, showed impaired kidney or liver function, or were affected by inflammatory bowel disease. Finally, 205 patients who met the study criteria were considered eligible for the evaluations.

Evaluations
Hematologic parameters (ALP, PSA) were evaluated before each Ra-223 administration. Bone scintigraphy to evaluate aBSI was performed before treatment and at its discontinuation, with the percentage change in those values calculated. The association of percentage changes in ALP and PSA with that of aBSI was evaluated to determine prognostic value. The Common Terminology Criteria for Adverse Events (CTCAE) package, version 4.03, was used for grading AEs [17]. The assessment of toxicity was performed at each treatment cycle.

Statistical Analysis
Spearman's rank correlation coefficient was used to assess the relationships of percentage changes in ALP and PSA with aBSI. Correlation strength was categorized using conventional statistical criteria, with 0-0.19 regarded as very weak, 0.2-0.39 as weak, 0.40-0.59 as moderate, 0.6-0.79 as strong, and 0.8-1 as very strong.
For evaluating the prognostic value of changes in ALP and PSA, and also aBSI, univariate and multivariate analyses were performed using a Cox proportional hazard model with stepwise regression. Any change in those parameters from the baseline was categorized as a decrease or increase/no change. To estimate survival distribution, the Kaplan-Meier product limit was used. A log-rank test was utilized for survival difference analysis.
Data analyses were performed using SPSS (version 21.0) and all tests were two-sided. p values < 0.05 were considered to indicate statistical significance.

Correlation of Changes in Serum Markers ALP and PSA with aBSI
aBSI change (%) showed a very weak but significant correlation with ALP change (r = 0.17, p = 0.017) and a weak but significant correlation with PSA change (r = 0.33, p < 0.0001) (Figure 2).

Survival Analysis
After a median period of 21.1 months (range 2.8-62.9 months, mean ± SD 24.2 ± 14.6 months) from the initial Ra-223 administration, 90 (43.9%) of the 205 patients died from prostate cancer. Univariate analysis showed that all analyzed parameters, including change in ALP (p = 0.028), PSA (p < 0.0001), and aBSI (p < 0.0001), were significant prognostic factors for OS (  A Kaplan-Meier curve was used, which demonstrated that the median OS for the 148 patients with ALP decline (median OS months) was significantly greater than that of the 57 patients with an increase or no change in ALP (22.6 vs. 15.8 months, p = 0.015), as shown in Figure 3a. Kaplan-Meier curve findings also demonstrated that the median OS of the 57 patients with PSA decline was significantly longer than that of the 148 patients with an increase or no change in PSA (32.5 vs. 16.8 months, p < 0.001), with those results shown in Figure 3b. Additionally, the median OS of the 108 patients with aBSI decline was significantly longer than that of the 97 patients with an aBSI increase or no change (27.1 vs. 16.2 months, p < 0.001) (Figure 3c). Finally, the median OS of the 43 patients with declines in both aBSI and PSA (33.3 months) was significantly longer (p < 0.001) than that of the 79 patients with a decline in only aBSI or PSA (20.5 months) and that of the 83 patients with aBSI increase/no change and PSA increase/no change (15.2 months) (Figure 3d).

Discussion
Although there are inherent limitations of a retrospective review of patient data, the present multi-center cohort of mCRPC cases that underwent Ra-223 therapy, and also preand post-treatment bone scintigraphy examinations is the largest such group reported, with the findings considered useful for analysis of the real-world impact of Ra-223. It was clarified that ALP change is very useful as a marker to evaluate Ra-223 treatment response, while PSA change is a very useful marker for prognosis prediction, and aBSI change is very useful for both evaluating treatment response and predicting the prognosis of treated patients. Furthermore, to the best of our knowledge, this study is the first to evaluate correlations of changes in three biomarkers (ALP, PSA, and aBSI). The results showed relatively low correlations among those, though they might be useful to indicate Ra-223 treatment response.

Discussion
Although there are inherent limitations of a retrospective review of patient data, the present multi-center cohort of mCRPC cases that underwent Ra-223 therapy, and also preand post-treatment bone scintigraphy examinations is the largest such group reported, with the findings considered useful for analysis of the real-world impact of Ra-223. It was clarified that ALP change is very useful as a marker to evaluate Ra-223 treatment response, while PSA change is a very useful marker for prognosis prediction, and aBSI change is very useful for both evaluating treatment response and predicting the prognosis of treated patients. Furthermore, to the best of our knowledge, this study is the first to evaluate correlations of changes in three biomarkers (ALP, PSA, and aBSI). The results showed relatively low correlations among those, though they might be useful to indicate Ra-223 treatment response.
Decreases in ALP, PSA, and aBSI were observed in 72% (148/205), 28% (57/205), and 53% (108/205), respectively, of the present Ra-223-treated cases, similar to data presented in the Anand study, which showed those reductions in 81% (54/67), 13% (9/67), and 36% (24/67), respectively [12]. Similarly, Prelaj et al. [18] reported that decreases in ALP and PSA following Ra-223 therapy were observed in 78% (25/32) and 59% (19/32), respectively, of their cases. In the ALSYMPCA trial, 47% showed an ALP response and 16% a PSA response (≥30% reduction from baseline) [2]. It has also been demonstrated that ALP, a biomarker of osteoblast activity, is superior to PSA as a biomarker of Ra-223 efficacy [10,[18][19][20]; thus, that has emerged as the leading biomarker in treated patients. An increase in PSA can be the result of the development of lymph node or visceral metastasis, which is not affected by Ra-223. Based on the results obtained in the present study and the natural target of Ra-223, the systemic bone scan parameter aBSI should be considered as the main parameter in association with ALP when evaluating response to radionuclide treatment.
Whether ALP change is an accurate biomarker for predicting survival remains controversial. One study demonstrated that an ALP response ≥30% or ≥10% reduction from the baseline value in patients receiving Ra-223 therapy can be a good predictor of OS [21], whereas three others found that ALP response after Ra-223 was not effective for OS prediction [10,12,19]. In another study, Sartor et al. [22] showed that ALP decline at 12 weeks after the first Ra-223 administration was correlated with OS, though the results did not meet the statistical requirements to indicate significance. However, the PSA level does not provide accurate information to elucidate the extent of skeletal metastasis or the effects of treatment on bone disease progression, though the recommended regimen does provide a survival benefit [23]. Moreover, Prelaj et al. [18] found that PSA response was significantly associated with survival, similar to the present findings, while two other groups [10,12] also reported univariate and multivariate analysis results showing that aBSI change was significantly useful for predicting survival.
When monitoring the effectiveness of therapy using assessments of clinical, biochemical (e.g., PSA, ALP), and imaging findings in prostate cancer patients treated with androgen deprivation therapy, cytotoxic chemotherapy, hormone therapy, and palliative radiotherapy, as well as Ra-223 therapy [24][25][26][27][28][29], care must be taken regarding the so-called "flare effect", which results from an early or transient rise in PSA level or tracer uptake during a bone scan procedure, followed by a later decline, especially in the early phase of treatment. Bone flare phenomena related to bone scintigraphy during the first three months of Ra-223 treatment have been reported [24][25][26]. Keizman et al. [26] also noted that a transient increase in bone metastasis-related pain was observed in 27% of their analyzed patients, while bone scintigraphy findings showed an increase in the number of bone lesions at three months when compared with the baseline in 26% of those cases and in 6% at six months as compared with three months. Castello et al. [27] reported PSA flare in 9.5-35.7% of prostate cancer patients treated with Ra-223, while better OS was noted in mCRPC patients experiencing PSA flare during Ra-233 therapy as compared to those with a progressive PSA increase. Following treatment, other imaging tools, such as fluorodeoxyglucose-positron emission tomography (FDG-PET), prostate-specific membrane antigen (PMSA)-PET, and fluoride-PET, can be useful to assess tumor load reduction, particularly in patients showing increased PSA, which is helpful for differential diagnosis between progression and pseudo-progression (related to PSA flare phenomenon) [27][28][29]. Based on their findings, Castello et al. [27] strongly recommended that Ra-223 therapy should not be discontinued after an early and transient PSA rise. In particular, physicians should be aware of the possibility of PSA flare induced by 233RaCl2 therapy during at least the first two months of treatment and that it does not represent a sign of disease progression.
A single-center study found that aBSI calculation derived using the DASciS software package, developed by the Sapienza University of Rome, is useful for predicting survival in mCRPC candidates treated with Ra-223 [11], which was also shown by the results of a multicenter study conducted in Italy [30]. This package has been demonstrated to be a simple tool that requires no more than a single demonstrative training session for a participating center. After gaining experience, the estimated time for a single BSI calculation is less than one minute, which is significantly lower than the time required for manual BSI calculation.
This study has some limitations, including its retrospective nature and inclusion of results obtained at multiple institutions. Therefore, generalization of the findings is limited and statistical errors are possible. Additionally, following the failure of Ra-223 therapy, the modalities used for the next chosen treatment were not uniform and various methods were used according to the choice of the attending physician. To more accurately clarify the usefulness of aBSI as an imaging biomarker for the evaluation of the response to Ra-223 therapy in clinical settings, a prospective multicenter trial that includes a larger cohort will be necessary.

Conclusions
The present findings indicate that the change in aBSI shown by bone scintigraphy is accurate and reliable for use as an imaging biomarker to assess treatment response in mCRPC patients treated with Ra-223, as well as to predict prognosis. Its use may provide for better management of mCRPC patients undergoing Ra-223 therapy.

Informed Consent Statement:
The requirement for written informed consent for study participation was waived by the institutional ethics committee. Instead, participants were given an opportunity to opt-out if they did not want their information to be used in this study.

Data Availability Statement:
The data presented in this study are available in this article.

Conflicts of Interest:
Author Kitajima has received research grants from PDRadiopharma, Inc. (Tokyo, Japan). The sponsor of the study had no role in design or conduct of the study, data collection or interpretation, or preparation of the report. There are no other potential conflicts of interest relevant to this study to be reported.