Is It Time to Revisit How We Measure and Report the Toxicities of Antineoplastic Pharmaceutical Agents?

A recent report examining the published descriptions of adverse effects (AEs) of therapy in phase 3 cancer trials serves as a poignant introduction to this commentary.1 In this review of 407 trials that collectively enrolled more than 300,000 patients, 16% failed to note the trial discontinuation rate due to AEs, only one-third fulfilled the criteria for complete toxicity reporting, and almost one-half (46%) met criteria for “subjective toxicity-minimizing language.”1 While recognizing that many study investigators quite appropriately presented relevant negative effects of the tested strategies, the cumulative data highlighted above suggest a clinically meaningful issue requiring further discussion.

In any dialogue of the status of toxicity reporting, it is relevant to highlight the National Cancer Institute’s Common Terminology Criteria for Adverse Events, which have been widely employed in clinical trials for several decades.2 While modifications have occurred over the years, it is essential to note that the reporting system itself, AE categories, and descriptions of criteria that define the degree of severity observed were initially established in an era that, for a number of reasons, was quite different from today.

During this earlier time period, the most concerning toxicities of the cytotoxic chemotherapeutic agents employed in routine oncology care were to the bone marrow (eg, neutropenia, anemia, thrombocytopenia) and gastrointestinal tract (eg, nausea, vomiting, diarrhea). Of course, there were other frequent but less medically serious effects (eg, hair loss) and rarer but potentially life-threatening effects to vital organ systems (eg, doxorubicin-induced heart failure, cisplatin-induced renal toxicity, bleomycin-induced pulmonary toxicity).

In most circumstances, these impacts were relatively easy to measure (eg, with blood counts, renal function tests, cardiac ejection fraction measurement) or directly observe (eg, emesis, degree of hair loss). However, this was not always the case (eg, fatigue). Even when the negative outcome was more subjective (eg, feeling of nausea), it was often possible to correlate the consequences with a related measured event (eg, number of vomiting episodes in the initial 24 hours after treatment). But perhaps most importantly, with rare exceptions (eg, cardiac drug toxicity leading to heart failure), even the most serious short-term objective toxicities (eg, severe emesis or neutropenia) were relatively quickly reversible, assuming they were not associated with a fatal event.

Treatment itself was, with the exception of hormonal maneuvers, generally delivered systemically and on an intermittent schedule (every 3 to 4 weeks) to permit normal organ function recovery. Again, with important exceptions, the duration of cytotoxic chemotherapy was uncommonly more than 4 to 6 cycles, due to the limited therapeutic effectiveness and often unacceptable acute or cumulative (or both) AEs.

Today, much has changed in the standard-of-care management of multiple malignant conditions. Oral therapy taken on a daily schedule (perhaps with a short interruption) is increasingly common, as are maintenance strategies where antineoplastic therapeutics are delivered for extended periods, frequently exceeding 1 to 2 years. In addition, novel agents, including molecularly targeted therapeutics, immunotherapy (eg, checkpoint inhibitors), CAR T therapy, and antibody-drug conjugates have strikingly different mechanisms of action as well as toxicity profiles compared with the single-agent and combination cytotoxic regimens that characterized systemic oncologic care in most settings in the not-so-distant-past.

So, with these rather substantial changes in antineoplastic drug therapeutics, the agents themselves, as well as the method and timing of delivery, the question to be posed is whether the existing long-established drug toxicity measurement schema, including detailed definitions of the degree of severity, remains an accurate representation of the symptoms and negative effects on normal function experienced by patients with cancer.

For example, is grade 1 or 2 nausea reported by a patient (as recorded by a clinical investigator or research nurse) that lasts for several days following carboplatin administered intravenously on an every 3-week schedule equivalent to grade 1 or 2 nausea a patient is noted to experience daily when taking a maintenance oral antineoplastic? Further, do the definitions of nausea grade that focus on eating habits, oral intake, potential weight loss, and dehydration accurately capture the impact of this symptom complex on the patient’s quality of life while experiencing these effects or the psychological impact of anticipat- ing they will regularly recur with each daily pill during a planned long-term oral maintenance therapeutic regimen?2

While mild, moderate, or severe bone marrow suppression can be defined by blood counts taken at regularly defined intervals with the data reasonably and meaningfully employed to compare the impact of different agents, drug dosages, and schedules, just how objective are attempted comparisons in the absence of relatively easily measurable outcomes?

In addition, such assessments are likely to be even more problematic if the toxicity evaluations are performed by different noncancer specialists based on their individual more-qualitative-rather-than-quantitative observations. For example, ophthalmologists asked to perform an examination for possible drug-associated ocular toxicity will approach the task with considerable variation in their prior experience, training, time constraints, and professional interest in the formal assessment of antineoplastic effects.

Finally, it is important to consider the actual clinical relevance of existing models for describing the severity of AEs of treatment. In this discussion it is worthy to note the metrics employed to define the severity of bone marrow suppression are well-founded and, for example, in the case of neutropenia were in the distant past (prior to the availability of potent broad-spectrum antibiotics and effective growth factors) shown to correlate with the risk of the development of serious/fatal infectious events. Similarly, the incidence of bleeding is recognized to be associated with the depth of measured platelet nadirs.

However, the documentation of symptomatic toxic events will be largely based on what the research physician/nurse records for individual items based on an interrogation of the patient, rather than a direct recording of patient-reported outcomes (PROs). As noted above, this long-established approach may be effective in the assignment of a severity grade but may fail to truly capture the overall impact of the event(s) on the patient’s quality of life. One may report, for example, that the level of fatigue is only “grade 1” (“relieved by rest”) but to the patient, this may be quite unacceptable, particularly if episodes occur frequently or necessitate several daily periods of rest.

Of considerable importance, there is an increasing appreciation of the relevance of PROs in the assessment of the overall clinical utility of antineoplastic therapeutics. However, faithful reporting of these results within primary study publications remains less than optimal.3

In conclusion, the question to be asked is whether it is time for the oncol- ogy establishment to fundamentally reassess how we measure and report the AEs of the antineoplastic therapeutics we routinely administer to our patients.

References

1. Miller AM, Passy AH, Sherry AD, et al. Incomplete toxicity reporting and use of toxicity-minimizing language in phase III oncology trials. JCO Oncol Pract. Published online January 8, 2025. doi:10.1200/ OP-24-00735
2. Common Terminology Criteria for Adverse Events (CTCAE) Version 5.0. US Department of Health and Human Services. November 27, 2017. Accessed June 27, 2025. https://ctep.cancer.gov/ protocoldevelopment/electronic_applications/docs/ctcae_v5_quick_reference_5x7.pdf
3. Blackstone EC, Efficace F, Meyerhardt JA, Abel GA. Timely reporting of patient-reported outcomes in cancer clinical trials: an ethical imperative. J Clin Oncol. 2025;43(10):1176-1179. doi:10.1200/ JCO-24-02021