📄 medicine

Illustrative Integration of Patient-Reported Outcomes Version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE) Data in Dose-Cohort Review for a Phase I Clinical Trial

This paper illustrates how integrating Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE) with clinician-graded CTCAE data in a Phase I acute myeloid leukemia trial provides a more comprehensive assessment of patient symptom burden and tolerability to inform dose-escalation decisions.

Original authors: Carolyn Mead-Harvey, Brie Noble, Claire Yee, Gina Mazza, Blake Langlais, Sheetal Patel, Peter Trask, Lauren Rogak, Ethan Basch, Amylou Dueck, Gita Thanarajasingam

Published 2026-07-02

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Carolyn Mead-Harvey, Brie Noble, Claire Yee, Gina Mazza, Blake Langlais, Sheetal Patel, Peter Trask, Lauren Rogak, Ethan Basch, Amylou Dueck, Gita Thanarajasingam

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a clinical trial for a new cancer drug as a high-stakes cooking competition. The chefs (doctors) are testing a new recipe (the drug) on a small group of tasters (patients) to see if it's safe enough to serve to more people.

In the past, the only way to judge the recipe was for the head chef to look at the tasters and say, "You look a bit pale," or "You seem to have a stomachache." This is like the CTCAE (Common Terminology Criteria for Adverse Events), where doctors grade side effects based on what they can see or measure.

However, the tasters themselves know exactly how the food tastes. They know if the spice is too hot, if the texture is weird, or if they feel nauseous even if they don't look sick. This is the PRO-CTCAE (Patient-Reported Outcomes version of the CTCAE), where patients fill out a digital menu card to tell the chefs exactly how they feel.

The Experiment

This paper is a "practice run" or a dress rehearsal for a Phase I trial involving patients with a specific type of blood cancer (acute myeloid leukemia). The researchers wanted to see what happens if the chefs and the tasters share their notes together during the first round of tasting (the first cycle of treatment).

They took data from just 4 patients in the first group and created a new kind of "scorecard" that combines the doctor's observations with the patient's diary.

What They Found (The Taste Test Results)

When they compared the two sets of notes, they found some interesting gaps:

The Chef Missed the Spice: In 75% of the patients, the doctors saw at least one "Grade 3" (severe) problem. But when they looked at the patients' own reports, 100% of the patients reported at least one symptom that felt like a "Grade 3" problem to them.
The Hidden Flavor: Every single patient had at least one symptom where their personal report was much worse than what the doctor wrote down.
- Example: One patient told the doctors, "My mouth is incredibly dry, it's a 3 out of 4 on the bad scale." The doctor, looking at the patient, wrote down "No dry mouth."
- Example: Another patient reported having diarrhea so often it was a "4 out of 4," but the doctor only graded it as a "2."

The New Scorecards

To make this data easier to read, the team invented two new ways to visualize the results:

The "Swimmer" Plot: Imagine a pool where each lane is a patient. The length of the bar shows how long a symptom lasted. The color of the bar shows what the doctor thought (CTCAE), while a circle floating above the bar shows what the patient felt (PRO-CTCAE). This lets you instantly see if the patient's circle is "higher" (worse) than the doctor's bar.
The "Butterfly" Plot: This is like a mirror image. On the left side, you see the doctor's grades; on the right, you see the patient's scores. It's designed to show the two sides of the same coin side-by-side.

The Main Takeaway

The paper concludes that looking at the doctor's notes and the patient's diary together gives a much fuller picture of the "flavor" of the drug.

The Analogy: If you only listen to the chef, you might think the soup is fine. But if you listen to the tasters, you realize it's too salty.
The Result: In this small group, the patients reported more suffering than the doctors recorded. By combining both sources, the "taste testers" (the trial team) get a richer, more complete story about whether the drug is tolerable.

A Note of Caution

The authors are very careful to say this is just a demonstration. They are showing how to present this data, not saying that this specific method should be used to make final decisions yet. They warn that because the group was so small (only 4 people), we can't make big statistical claims. It's like judging a whole restaurant menu based on tasting just four bites; it's a great start, but you need more tasters to be sure.

In short, this paper is a blueprint for a new way of talking to patients during early drug trials, ensuring that the "silent" symptoms they feel aren't left out of the safety conversation.

Technical Summary: Illustrative Integration of PRO-CTCAE Data in Dose-Cohort Review for a Phase I Clinical Trial

Problem Statement
Phase I clinical trials are primarily designed to evaluate the safety and tolerability of new treatments. Historically, safety assessments rely on clinician-graded Adverse Events (AEs) using the Common Terminology Criteria for Adverse Events (CTCAE). However, prior research indicates that clinicians may miss or underreport symptomatic AEs. While Patient-Reported Outcomes (PROs) are standard in later-phase trials, there is no consensus on how to incorporate PRO data, specifically the Patient-Reported Outcomes version of the CTCAE (PRO-CTCAE), into the decision-making process for dose escalation in Phase I trials. Experts disagree on whether PROs should be used descriptively or for formal decision-making, and there is a lack of standardized methods for visualizing and summarizing PRO-CTCAE data alongside CTCAE data during interim dose-cohort reviews.

Methodology
This study is a secondary analysis of data from a Phase I clinical trial involving patients with relapsed or refractory acute myeloid leukemia (NCT02670044). The authors developed an illustrative framework for summarizing and visualizing safety data for the first dose-level cohort (N=4) during Cycle 1.

Data Collection: Patients completed PRO-CTCAE questionnaires regarding 24 selected symptoms on electronic devices at baseline (Day 1) and on Days 8, 15, and 22 of the first 28-day cycle.
Scoring and Adjustment: PRO-CTCAE attributes (frequency, severity, interference) were scaled 0–4. Composite scores were calculated by combining attributes into a single score scaled 0–3 to mimic CTCAE grading. To isolate treatment-emergent symptoms, a baseline adjustment was applied: the worst score during Cycle 1 was reported only if it exceeded the baseline score; otherwise, a score of zero was assigned.
Visualization and Reporting: The authors created a multi-modal summary for the cohort:
- Narrative Summaries: Patient-level narratives highlighting post-baseline PRO-CTCAE scores $\ge$ 3 and all-grade CTCAE events.
- Tabular Summaries: Cohort-level counts and percentages of patients with baseline-adjusted PRO-CTCAE scores > 0 and $\ge$ 3.
- Graphical Visualizations: "Swimmer" and "Butterfly" plots were utilized to display longitudinal data for specific symptoms (vomiting, nausea, diarrhea). These plots simultaneously display CTCAE grades (via bar color and length) and PRO-CTCAE composite scores (via markers), allowing for direct comparison of clinician and patient perspectives within the same timeframe.

Key Results
The analysis of the first dose cohort (4 patients) yielded the following findings:

Discrepancy in Severity: 3/4 (75%) of patients experienced at least one CTCAE Grade 3+ adverse event. However, 100% of patients had at least one PRO-CTCAE composite score of 3 during Cycle 1.
Underreporting by Clinicians: All 4 patients (100%) had at least one symptom where the PRO-CTCAE indicated a greater symptom burden than reflected in the CTCAE documentation.
- Example: In one patient, PRO-CTCAE reported dry mouth severity as 3 (baseline 0) and diarrhea frequency as 4 (baseline 2), whereas CTCAE recorded no dry mouth AE and only Grade 2 diarrhea.
Cohort-Level Prevalence: High prevalence of treatment-emergent symptoms was observed. For instance, 75% of patients had baseline-adjusted diarrhea frequency scores $\ge$ 3, and 100% reported nausea frequency and severity scores > 0.
Visualization Utility: The swimmer and butterfly plots effectively highlighted differences in reporter, scale structure, and assessment timing. For example, in one patient, nausea was graded as CTCAE Grade 1 (Days 4–8), while the PRO-CTCAE composite score was 2 on Day 8 and increased to 3 thereafter.

Significance and Claims
The paper positions this work as an illustrative example rather than a definitive guide for regulatory decision-making. The authors claim that:

Complementary Value: PRO-CTCAE data complements CTCAE reporting by capturing symptomatic burden from the patient perspective that is not always fully reflected in clinician-graded AEs.
Enriched Discussion: Reviewing PRO-CTCAE and CTCAE data together during dose-cohort evaluation may enrich discussions regarding safety and tolerability in Phase I trials.
Alignment with FDA Guidance: The approach aligns with FDA Project Optimus goals of dose optimization by considering safety and tolerability alongside efficacy, and adheres to FDA guidance recommending frequent (e.g., weekly) assessments during initial treatment cycles.
Limitations and Cautions: The authors explicitly state that this is a descriptive approach with a small sample size (N=4). They caution against overinterpreting results from small cohorts for formal hypothesis testing or advanced modeling. Furthermore, they note that while this summary enriches discussion, prospective evaluation is required before such summaries can be used to influence formal dose-escalation decisions. The paper does not claim that PRO-CTCAE should currently replace CTCAE for decision-making, but rather that the integration of both data sources offers a more comprehensive view of tolerability.