Arsenic and all-trans retinoic acid for acute promyelocytic leukemia: yes, it really is as good as it seems.

In this issue of Haematologica, Kayser and colleagues report the results of an analysis of outcomes from the National Acute Promyelocytic Leukemia (APL) Observational study (NAPOLEON-Registry; NCT02192619), including 152 non-high-risk APL patients in Germany and France.[1] In their study, which they claim represents a reflection of “real-life” outcomes, these authors specifically focused on APL patients treated upfront with all-trans retinoic acid (ATRA) and arsenic, according to the study led by the late Francesco Lo Coco.[2] As with that original protocol, this present study excluded high-risk APL patients.

When Lo-Coco’s study was published in 2013, the results seemed almost too good to be true.[2] The eventfree survival rate of patients treated with ATRA and arsenic was 97%. In their study of the registry patients, Kayser and colleagues found an almost identical result (event-free survival of 95%, with no patient relapsing after achieving remission. The remarkable efficacy of this regimen seems to be every bit as high even outside of the context of a clinical trial. Two out of 152 patients died during induction, and both were older (64 and 70 years) than typical APL patients. Interestingly, differentiation syndrome was only reported in seven patients (6%), in contrast to the 19% reported in Lo-Coco’s study. One wonders whether this is more a reflection of clinicians’ comfort in managing and even preventing this condition as they grow more familiar with this regimen over time.

Where to next with APL? Certainly, an oral version of arsenic would expand the use of this combination to many parts of the world lacking access to intravenous medication. It would also represent a major improvement in the quality of life of APL patients, who must trudge through months of daily arsenic infusions. Oral preparations are under investigation,[3] but formulation challenges have thus far been an obstacle to their widespread use.

High-risk APL patients were excluded from these studies, and of course they represent a significant challenge for physicians treating them. In one of the original pilot studies exploring the combination of ATRA and arsenic, gemtuzumab ozogamycin (GO) was used as a cytoreductive agent in the high-risk patients.[4] This highly effective agent is not approved for such use worldwide, but studies to compare its efficacy against anthracyclines are warranted.

Another way to potentially optimize this therapy is to determine how much arsenic is really needed to achieve these high-quality outcomes. The selection of four cycles of consolidation with arsenic was somewhat arbitrary, and no one should lose sight of the fact that arsenic is a group 1 human carcinogen with neurotoxic and hepatotoxic effects.[5] Identifying the minimum necessary number of cycles would be a worthwhile endeavor for the field.

Finally, lest we be too self-congratulatory about how well we are doing with this dreadful malignancy, let us not forget how many patients die of APL before their disease is recognized and treated.[6] At present, in areas of the world that have complete access to standard-of-care leukemia treatment, most APL patients die because their care providers are unknowingly observing the natural history of untreated APL. The failure to recognize APL rapidly is a problem without an immediate solution. However, perhaps in this digital age, there is a ray of hope for this problem. The use of artificial intelligence algorithms combined with digital scanning technology may offer an automated way of identifying an APL patient,[7] leading to the same sort of electronic red flag that occurs when a patient with a low electrolyte or platelet count is evaluated by an emergency room physician. We are probably not far off from that future.