It is estimated that in year 2013, approximately 238,590 men will be diagnosed and 29,720 will die due to prostate cancer (PC) in the United States (1). Despite the effectiveness of hormone therapy, every patient with metastatic disease is currently incurable and those who live long enough eventually progress to castration-resistant prostate cancer (CRPC) (2).

Radioimmunotherapy (RIT), using specific monoclonal antibodies (mAbs) or fragments which are radiolabeled (most typically with beta-emitting particles), has proven quite effective in the treatment of non-Hodgkin lymphoma (NHL) alone or in combination with chemotherapy (3–5). RIT for solid tumors has posed a more difficult challenge for a number of biologic, technical, and practical reasons (6). In the last decade, significant progress has been made in the development of RIT for solid tumors. As with any therapy, emphasis is laid on balancing toxicity and therapeutic effects.

Metastatic PC is a good candidate for RIT, because it is radio-responsive and typically develops as small-volume metastatic sites of disease in marrow and lymph nodes that receive high levels of circulating antibody. Several clinical trials have focused on or included subjects with PC (7–16). Importantly, unlike some other solid tumors, a well-established, specific cell-surface antigen has been identified: prostate-specific membrane antigen (PSMA) (17–19). PSMA is an ideal target as it is expressed by nearly all PCs and is not secreted (17, 19). The expression levels progressively increase in more poorly differentiated, metastatic, and castration-resistant cancers (17, 20).

J591 is a deimmunized mAb which specifically binds with high affinity to the extracellular domain of PSMA (21, 22). In addition, the PSMA-J591 antibody complex is internalized thereby delivering any radionuclide or drug conjugated to the antibody to the interior of the targeted cancer cells (23). We have performed and reported 4 clinical trials using ¹⁷⁷Lu and ⁹⁰Y labeled J591 (24–28) (Table 1). Based on imaging studies, we have shown that J591 is able to sensitively and specifically target sites of metastatic PC in both bone and soft tissue (28). As often seen with RIT, radiolabeled-J591 was well-tolerated and serious toxicity was confined to predictable, reversible myelosuppression (24, 25, 27, 28). The preliminary efficacy data from initial phase I studies has been confirmed in a phase II clinical trial of single-dose ¹⁷⁷Lu-J591 (28).

Several RIT studies have demonstrated that, in the absence of bone marrow or hematopoietic stem cell support, radiation-induced myelotoxicity is the dose-limiting toxicity (DLT) (29, 30). Hematologic toxicity is the biggest challenge faced by all radioimmunotherapeutic agents. The manifestations of myelotoxicity may be related to the pretreatment peripheral blood cell counts and bone marrow reserve, which may have been compromised by prior therapies (29, 31). While the vast majority of patients have spontaneous recovery of blood counts, some worry about the ability to deliver subsequent cytotoxic chemotherapy if needed clinically. The development of secondary myelodysplastic syndrome (MDS) or acute myelogenous leukemia (AML) has been reported (32) and has been estimated to be approximately 2–5% (33). Two studies have attempted to comprehensively evaluate the risk for MDS and AML following RIT with either ⁹⁰Y-ibritumomab or ¹³¹I-tositumomab. These retrospective analyses of large numbers of patients both showed that RIT did not demonstrate a higher risk for MDS in comparison with similar patient populations treated with multiple chemotherapies alone (34, 35).

Long-term outcomes have rarely been reported with solid tumor RIT. Here, we report the long-term toxicity data in patients treated with radiolabeled-J591, including bone marrow recovery and ability to deliver cytotoxic chemotherapy.

Between October 2000 and August 2012, 150 subjects with metastatic CRPC received radiolabeled-J591. One hundred and twenty-one patients received ¹⁷⁷Lu-J591 at total doses of 20–90 mCi/m² and 29 patients received ⁹⁰Y-J591 at 5–20 mCi/m². Patient baseline demographics are displayed in Table 2.

Prostate cancer offers a model for the investigation and development of RIT in solid tumors, given the restricted, high-level of expression of PSMA and the availability of a specific, well-tolerated mAb (J591). As myelosuppression and subsequent bone marrow recovery as well as the theoretical inability to tolerate subsequent cytotoxic chemotherapy are potential issues with RIT, we analyzed our long-term results with radiolabeled-J591. It should be noted that this analysis included radiolabeled-J591 administered across a number of different cumulative doses in phase I and II studies, including a prospective fractionated dose schedule (Table 1).

The vast majority of subjects who received radiolabeled-J591 had complete (i.e., grade 0) or near-complete (grade 1) recovery of neutrophil (100%) and platelet (97.3%) counts. Those that did not experience complete recovery or who experienced subsequent decline in blood counts had concomitant progression of PC. Those with bone marrow biopsies at subsequent count decline had diffuse PC marrow infiltration and none had evidence of MDS or leukemia. As some have proposed that prior treatment might influence subsequent bone marrow reserve and the ability to tolerate RIT, (31) we analyzed by previous exposure to chemotherapy and/or radiation. There was a trend for more grade 4 thrombocytopenia in those with both prior chemotherapy and radiation, but no clear differences in those who had received either previous chemotherapy or radiation.

Because of the myelosuppression observed in patients treated with radiolabeled-J591, a further concern is whether patients can tolerate subsequent therapeutic interventions if needed for disease progression after RIT. The available data suggest that patients treated with radiolabeled-J591 can tolerate subsequent therapies. The majority (92 of 150, 61.3%) of patients received cytotoxic chemotherapy after radiolabeled-J591, including docetaxel and cabazitaxel. Although two of the four studies were completed prior to the approval of docetaxel for metastatic CRPC, 95 of 150 (63.3%) received docetaxel at any time, a favorable number compared to the estimated third of men with metastatic CRPC ever receiving docetaxel according to a survey of urologists and oncologists in the era prior to the approval of newer agents such as Sipuleucel-T, Abiraterone, and Enzalutamide. Another concern might be a theoretic possibility of poorer chemotherapy tolerance in the setting of prior RIT. While this cannot be completely addressed in a retrospective study, patients who received docetaxel at any time following radiolabeled-J591 received a median of 23 weeks of chemotherapy (range 10–68) (i.e., 7–8 cycles) and those who received cabazitaxel received a median of 18 weeks (range 17–36) (i.e., 5–6 cycles), both within range of general community standards.

One of the most serious concerns about RIT is its potential to lead to secondary malignancies, particularly MDS/AML. The long-term data following RIT of solid tumors is limited, in part for the practical reason of populations that have been studied (generally late stage, refractory cancers). In this study, subjects were treated over a 12-year period with median time from treatment of 7.8 years (range 0.7–12.7); censored for loss to follow-up or death, median follow-up for this analysis was 16.6 months (range 0.5–133.9). The estimated median survival for the population based upon nomogram analysis is 15 months (median Halabi score 145, range 62–196) (36). As solid tumor RIT moves earlier in the disease process and as median survival for advanced solid tumors increases with more successful therapy, long-term toxicity follow-up is needed. More data exist following RIT for NHL. With a median 10-year follow-up, Kaminski et al. reported a single case of MDS out of 76 previously untreated patients who received a single treatment with ¹³¹I-tositumomab (37). Czuczman et al. reviewed the records of 746 patients treated with ⁹⁰Y-ibritumomab tiuxetan with a median follow-up of 4.5 years and found a total of 2.5% of patients developed MDS at a median of nearly 2 years following RIT (35). This corresponded to an annualized MDS rate of 0.7% per year following RIT. The expected annual rate of treatment-related MDS for NHL patients receiving systemic therapy alone or in combination with rituximab is approximately 1% per year (38, 39). Higher doses of therapy may pose greater risk. Recently, Guidetti et al. published prospective data on development of MDS/AML in NHL patients receiving myeloablative doses of ⁹⁰Y-Ibritumomab tiuxetan. Among 52 patients with a median follow-up of 49 months, the 5-year cumulative incidence of MDS/AML was 8.29% (40). This incidence is significantly higher than previously reported with lower doses, but the studied population consisted of subjects receiving significantly higher doses of RIT than used for solid tumors and a matched-pair analysis of patients receiving myeloablative chemotherapy conditioning instead of RIT revealed a similar 8.05% 5-year cumulative incidence of MDS/AML.

Taken together, although there have been few prospective studies designed specifically to assess the risk of development of MDS/AML following RIT for NHL, with significant follow-up, the use of RIT does not appear to significantly increase the risk of secondary MDS above the risk of chemotherapy alone in patients with NHL. Should cases of MDS or leukemia be discovered with radiolabeled-J591 or other RIT for PC, it will be important to understand that as in NHL, additional therapies may be associated with MDS/AML and there is a de novo incidence in an elderly male population (35, 41, 42).

Systemic targeted radiation with RIT has therapeutic promise in advanced solid tumors, in particular for radiosensitive tumors such as PC with a selective and specific cell-surface antigen such as PSMA and an available antigen-specific, non-immunogenic mAb such as J591. Using beta-emitting radionuclides at therapeutic doses, myelosuppression is expected. However, toxicity is predictable and usually self-limited, with the majority of patients who do not refuse able to receive chemotherapy. As the use of RIT becomes more attractive across diseases with the development of newer, more specific mAbs or peptides, studies examining long-term toxicity are warranted.

Neil H. Bander is an inventor on patents that are assigned to Cornell Research Foundation (“CRF”) for the J591 antibody described in this article. Dr. Bander is a paid consultant to and owns stock in BZL Biologics, the company to which the patents were licensed by CRF for further research and development. The other co-authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.