Is the pendulum slowly starting to swing back?

Concerns related to the cardiac effects of trastuzumab go back almost as far as the discovery of the agent that was first approved for use in the United States in 1998. An early report, derived from data from patients who had received the agent concomitantly with anthracycline therapy noted that 27% of treated patients showed markers or symptoms of cardiotoxicity [1]. At the time trastuzumab was being considered for use in the adjuvant setting because of broad efficacy in treating HER2 positive breast cancer, yet concerns were sufficiently strong that a warning was published, noting that “…aggressive use of this agent could test the time-worn dictum that physicians should do no harm” [2]. Trastuzumab was approved for adjuvant use and made a huge improvement in treating a form of breast cancer that had previously been felt to have a less favorable prognosis. Concerns regarding cardiotoxicity continue to be an issue; the variance in the reported incidence suggests that much information related to this agent is uncertain; reports sometimes still mention the initially reported 27% incidence, although modern use, even in the face of long-term follow-up, seems not to approach that level.

At that time when adjuvant approval was being considered, the cardiac effects of anti-cancer therapy were thought to be a single entity. The negative effect of anthracyclines on cardiac myocytes was well known, and the perception that what was seen with trastuzumab was similar was not surprising. Based on what was known at that time our group speculated that what we were observing might be direct toxicity, a synergistic or sequential stress phenomenon, or a surveillance artifact [3]. Cardiotoxicity was, and to some extent still is considered to be a huge concern. Trastuzumab is generally not given concomitantly with anthracyclines; current usage generally avoids the synergistic effect. When used sequentially the late effects of anthracyclines may greatly confuse the etiology of subsequent cardiac dysfunction [4]. The concerns of trastuzumab-associated cardiotoxicity, while less of an issue today, remain an ongoing consideration.

Among the steps taken to define the extent and impact of trastuzumab cardiotoxicity was the recognition that whatever the drug was doing, it was clearly different from the predictable dose-related toxicity of the anthracyclines. The description of that difference was widely cited, but not without initial and ongoing criticism [5, 6]. With the publication of several clinical trials, it was noted that the level of reported toxicity was often in a narrow range, below 5%. Our group performed a modeling exercise, whereby the likelihood of false positive cardiac ultrasound was undertaken; the result, 3.6% was remarkably consistent with some of the reports from these clinical trials [7]. The earlier posed speculation as to the extent of surveillance artifact was finally entering into mainstream consideration [8].

The present monitoring recommendation for patients being treated with trastuzumab, as published as the 2022 European Society of Cardiology guidelines, is that they be monitored every three months during the year of therapy [9]. Among the questions to be considered are: (1) is the present scheme of monitoring recommendations still one that should be applied, and (2) does interruption of treatment in cases deemed to indicate toxicity cause greater good than harm? We must ask if the present strategies address the balance between cardiac risk and oncologic benefit optimally.

In a recent paper published in Cardio-Oncology, intriguing evidence is provided that leads us a step closer to recognizing these questions. Dina Labib et al. present data from a long-term cardiac follow-up of the MANTICORE (Multidisciplinary Approach to Novel Therapies in Cardio-Oncology) trial. MANTICORE initially showed the potentially beneficial effect of beta-adrenergic blockers and angiotensin-converting enzymes inhibitors during a year of trastuzumab treatment. In the present report, 52 (of the original 94) patients were studied at a median of 6.5 years post randomization to the original treatment arms. The finding that the group that did not receive interventional drugs exhibited recovery to the same extent as those on the placebo arm, suggests that long-term follow-up may not be as important as previously considered [10]. The authors note that surveillance recommendations vary as to follow-up beyond one year, but hypothesize that long-term surveillance may not have added benefit, even acknowledging that left ventricular mass has declined. The mechanism of the decline has not been definitely determined and could well be multifactorial.

The reported update, unfortunately, has a limited number of subjects for evaluation and relies on volunteers for continual treatment begging the question of selection bias. Hence it cannot provide a final answer to the question. A robust answer is unlikely to be forthcoming as it will need larger groups of patients that can be followed for long periods of time; separating the late effects of anthracyclines from those of trastuzumab will remain challenging even when administered sequentially. The authors’ conclusion, that “[l]onger term surveillance may be warranted in some high-risk breast cancer populations…” seems warranted, but new information needs to be incorporated in our practice guidelines as it becomes available. The authors note that “[t]he significance of altered cardiac geometry along the breast cancer treatment and survivorship trajectory warrants further study” seems on target as well. While this paper does not directly answer either of the cardinal questions, it moves us a small step further in our goal to establish meaningful guidelines that balance cost, inconvenience, and both long-term and short-term balance of risks and benefits to optimizes true and meaningful advantages for our patients. Oncologists do not strictly follow present guidelines for cardiac monitoring [11]. While many reasons may be given, one may well be that they feel the guidelines may be overly inclusive. Their position may reflect that the pendulum is starting to move back from its present position. It is papers like the present one by Dina Labib et al. that may help us ultimately find the optimal monitoring and intervention schedules. It is refreshing to see the pendulum move; hopefully it is moving in the right direction.