Prediction of Moderate to Severe Delayed Chemotherapy‐Induced Nausea Using a Novel Algorithm Incorporating Individual Patient Factors.

Prediction of Moderate to Severe Delayed Chemotherapy‐Induced Nausea Using a Novel Algorithm Incorporating Individual Patient Factors - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Cancer Med . 2026 Apr 17;15(4):e71824. doi: 10.1002/cam4.71824 Search in PMC Search in PubMed View in NLM Catalog Add to search Prediction of Moderate to Severe Delayed Chemotherapy‐Induced Nausea Using a Novel Algorithm Incorporating Individual Patient Factors Alisha P Maity Alisha P Maity 1 Lankenau Medical Center, Alliance Cancer Specialists, Wynnewood, Pennsylvania, USA Find articles by Alisha P Maity 1 , Stephanie D Kjelstrom Stephanie D Kjelstrom 2 Lankenau Institute for Medical Research, Wynnewood, Pennsylvania, USA Find articles by Stephanie D Kjelstrom 2 , Kathryn J Fleck Kathryn J Fleck 2 Lankenau Institute for Medical Research, Wynnewood, Pennsylvania, USA Find articles by Kathryn J Fleck 2 , Erik L Zeger Erik L Zeger 1 Lankenau Medical Center, Alliance Cancer Specialists, Wynnewood, Pennsylvania, USA 2 Lankenau Institute for Medical Research, Wynnewood, Pennsylvania, USA Find articles by Erik L Zeger 1, 2 , Zonera A Ali Zonera A Ali 1 Lankenau Medical Center, Alliance Cancer Specialists, Wynnewood, Pennsylvania, USA 2 Lankenau Institute for Medical Research, Wynnewood, Pennsylvania, USA Find articles by Zonera A Ali 1, 2 , Arezoo Ghaneie Arezoo Ghaneie 1 Lankenau Medical Center, Alliance Cancer Specialists, Wynnewood, Pennsylvania, USA 2 Lankenau Institute for Medical Research, Wynnewood, Pennsylvania, USA Find articles by Arezoo Ghaneie 1, 2 , Aarti L Shevade Aarti L Shevade 1 Lankenau Medical Center, Alliance Cancer Specialists, Wynnewood, Pennsylvania, USA Find articles by Aarti L Shevade 1 , Paul B Gilman Paul B Gilman 1 Lankenau Medical Center, Alliance Cancer Specialists, Wynnewood, Pennsylvania, USA 2 Lankenau Institute for Medical Research, Wynnewood, Pennsylvania, USA Find articles by Paul B Gilman 1, 2 , Deric C Savior Deric C Savior 1 Lankenau Medical Center, Alliance Cancer Specialists, Wynnewood, Pennsylvania, USA 2 Lankenau Institute for Medical Research, Wynnewood, Pennsylvania, USA Find articles by Deric C Savior 1, 2 , U Margaretha Wallon U Margaretha Wallon 2 Lankenau Institute for Medical Research, Wynnewood, Pennsylvania, USA Find articles by U Margaretha Wallon 2, ✉ Author information Article notes Copyright and License information 1 Lankenau Medical Center, Alliance Cancer Specialists, Wynnewood, Pennsylvania, USA 2 Lankenau Institute for Medical Research, Wynnewood, Pennsylvania, USA * Correspondence: U. Margaretha Wallon ( wallonm@mlhs.org ) ✉ Corresponding author. Revised 2026 Mar 30; Received 2025 Nov 4; Accepted 2026 Apr 3; Collection date 2026 Apr. © 2026 The Author(s). Cancer Medicine published by John Wiley & Sons Ltd. This is an open access article under the terms of the http://creativecommons.org/licenses/by-nc-nd/4.0/ License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non‐commercial and no modifications or adaptations are made. PMC Copyright notice PMCID: PMC13090175  PMID: 41998835 ABSTRACT Purpose Chemotherapy‐induced vomiting and nausea (CINV) is difficult to accurately predict, especially in the delayed phase and can negatively affect a patient's' quality of life as well as reduce their motivation to continue lifesaving therapy. We have previously reported on a blood‐based assay that can be used to tailor an individualized antiemetic prophylaxis. The purpose of this current study was to validate our previous finding in a larger cohort of patients treated with a platinum‐containing regimen. Methods Blood samples and patient reported outcomes from 202 cancer patients aged 18 years and older who started treatment with a platinum agent during February 15, 2016, through May 18, 2023, were analyzed. The glutathione recycling capacity in red blood cells was compared to reported nausea during the delayed phase and verified by comparison to notes in medical records. An algorithm for risk score prediction was generated and predicted risk scores were compared to observed delayed nausea severity. Results Patients with a low glutathione recycling capacity, indicating prolonged exposure to oxidative stress, were experiencing moderate/severe delayed chemotherapy‐induced delayed nausea more than those who had an efficient recycling capacity to offset the excess radical oxygen formed by chemotherapy. Calculated risk scores from participating patients demonstrated a good correlation between predicted and observed outcomes for the risk score groups. Conclusion Risk score, based on glutathione recycling and other intrinsic factors, can predict the risk of delayed nausea for patients undergoing chemotherapy, allowing for a more aggressive preventive treatment upfront to avoid or reduce delayed nausea symptoms. Keywords: chemotherapy‐induced nausea and vomiting, delayed nausea, oxidative stress, platinum agents, risk prediction 1. Introduction Chemotherapy‐induced nausea and vomiting (CINV) is a common and distressing side effect for oncology patients [ 1 , 2 , 3 ]. Nausea and vomiting in these patients can be acute (≤ 24 h after treatment) or delayed (> 24 h after treatment) [ 4 ]. Unaddressed CINV affects patients' abilities to tolerate chemotherapy and can lead to worsening quality of life and performance status, anticipatory nausea, and result in dose reductions or early termination of chemotherapy [ 5 ]. The management of CINV is complex, as it is impacted by both the emetogenicity of the chosen chemotherapeutic agent and the occurrence of two distinct pathophysiological phases (acute and delayed) which can result in different outcomes and require distinctive treatment approaches [ 4 , 6 ]. Therefore, organizations such as the Multinational Association of Supportive Care in Cancer (MASCC) [ 7 , 8 ], the American Society of Clinical Oncology (ASCO) [ 9 ], and the National Comprehensive Cancer Network (NCCN) [ 10 ] have generated guidelines to aid clinicians in optimal prophylaxis of CINV. These guidelines recommend that when oncologic treatments are chosen, the emetogenic potential of chemotherapeutic agents should be assessed so that providers can prescribe appropriate prophylactic antiemetic agents [ 10 , 11 ]. Even with these guidelines, CINV remains an issue for many patients, prompting numerous investigations into potential risk factors for CINV, with the goal of determining a prophylactic strategy for medications. These factors include history of CINV with prior chemotherapy and/or perceived susceptibility to nausea, female sex, low chronic alcohol intake, and younger age [ 10 , 11 ] and have been used to assess patients' risk for the development of CINV [ 12 ]. However, the patients' individual biology is not included in previous predictive algorithms. The pathophysiology of CINV, and nausea especially, is not well understood, though one proposed mechanism suggests that CINV is triggered by free radicals in the small intestine induced by chemotherapy. These free radicals, such as reactive oxygen species (ROS), cause release of serotonin and substance P from enterochromaffin cells [ 13 , 14 ], subsequently activating their respective receptors and inducing nausea and vomiting [ 15 ]. Based on this mechanism, we have previously developed a hypothesis that a patient's risk of CINV is directly related to their efficiency in scavenging free radicals via glutathione (GSH), a major antioxidant, after exposure to chemotherapy. Specifically, we have focused on measuring recycling of GSH (GSHrc) as a mechanism by which cells detoxify the reactive oxygen species released after chemotherapy [ 16 , 17 ]. In brief, the method is based on enzymatic conversion of the bioactive probe hydroxyethyl‐disulfide (HEDS). Due to reported single‐nucleotide polymorphisms among the enzymes in the recycling process [ 18 ], individual variation in the rate of GSHrc, and thereby detoxification, can affect ROS levels and in extension release of neurotransmitters which activate the emetic response [ 15 , 19 ]. We have previously described preliminary results for our development of the glutathione recycling (GSHrc) assay and outlined its relationship to CINV [ 20 , 21 ]. We developed a predictive model for chemotherapy‐induced nausea and vomiting (CINV) using data from 202 patients who received platinum‐based chemotherapy for colorectal, lung, esophageal, head and neck, and other non–hormone‐driven malignancies. These cancer types were selected to ensure balanced representation by sex and inclusion of all three commonly used platinum agents—carboplatin, cisplatin, and oxaliplatin. Our prospective analysis led to the creation of a risk score to estimate individual CINV risk. This model may support clinicians in making personalized decisions for the prevention and management of CINV. 1.1. Study Objectives The primary objective of this study was to develop a predictive model for chemotherapy‐induced nausea and vomiting (CINV) that integrates both clinical risk factors and a patient‐specific biological measure of glutathione recycling (GSHrc). We aimed to generate a practical risk score that can stratify patients by their likelihood of developing CINV after platinum‐based chemotherapy. With the larger cohort, we could utilize machine learning techniques, allowing for the inclusion of more variables to improve personalized prophylaxis and optimize supportive care for patients undergoing chemotherapy. 2. Methods 2.1. Patient Cohort Our prospective observational study was performed as a single‐center investigation conducted at Lankenau Medical Center through a collaboration between the Alliance Cancer Specialists oncology clinic and the Lankenau Institute for Medical Research (LIMR). A total of 428 patients with a cancer diagnosis who initiated treatment with chemotherapy between February 15, 2016, and May 18, 2023, were enrolled in the study as approved by the local Institutional Review Board (IRB) (Reference ID: F/N‐R12‐3146 L). Applying exclusion criteria (Figure 1 ) and removing patients lost to follow‐up resulted in a cohort of 202 patients with analyzable data. Patients across all socioeconomic groups, races, and ethnicities were encouraged to participate. However, the population in our health system area is predominantly Black and White, with approximately 10% identifying as Hispanic, Asian, or other. Written and signed informed consent was obtained from all patients. Adults (> 18 years old) of both sexes were permitted following the protocol, and all patients had to be chemotherapy naive. Patients with a recurrence or a second primary were not enrolled in this study. All patients with a solid tumor and receiving a platinum‐based therapy were treated according to current medical standards; the medical staff had exclusive authority in recommending therapy based on stage of disease and/or co‐morbidities. Patients treated with a highly emetogenic chemotherapy (HEC) [ 5 , 10 ] including cisplatin, were given 0.25 mg palonosetron IV, aprepitant 150 mg IV, and 10 or 20 mg dexamethasone IV prior to chemotherapy (all antiemetic regimens given are described in Table 1 ). Most patients treated with a carboplatin‐containing regimen (85%) received the agent at Area Under the Curve (AUC) 2, which is considered a moderately emetogenic chemotherapy (MEC) [ 10 ] and given with a 2‐drug antiemetic preventive treatment (0.25 mg palonosetron IV and 10 or 20 mg dexamethasone IV) [ 10 , 22 ]. Oxaliplatin was given at either 85 mg/m 2 or 130 mg/m 2 depending on the regimen. This patient group received the same antiemetic regimen as patients treated with carboplatin AUC 2. All patients were given prescriptions to fill for ondansetron and prochlorperazine if treated with oxaliplatin or carboplatin (AUC < 4). Patients treated with cisplatin or carboplatin (AUC ≥ 4) were prescribed the same agents with the possible addition of olanzapine. All prescriptions were for pro re nata (PRN) use during the 24–96 h post‐treatment, and instructions were given by the medical oncology staff for dosing frequency. FIGURE 1. Open in a new tab Consort diagram of the study. Of the 428 enrolled patients, 290 were diagnosed with solid tumors but not breast malignancy, received a platinum‐containing therapy, and did not report nausea and/or vomiting immediately before treatment. Patients lost to follow up due to death or relocation were also excluded, resulting in a cohort of 202 patients with analyzable data for their treatment regimen. TABLE 1. Chemotherapy and antiemetic protocols. Chemotherapy Emetogenicity Dose/Regimen Antiemetic prophylaxis PRN medications Cisplatin High (HEC) Standard Palonosetron 0.25 mg IV, aprepitant 150 mg IV, olanzapine 10 or 5 mg, dexamethasone 10–20 mg IV Ondansetron, prochlorperazine, ± olanzapine Carboplatin Moderate (AUC 2) AUC 2 Palonosetron 0.25 mg IV, dexamethasone 10–20 mg IV Ondansetron, prochlorperazine, ± olanzapine if AUC ≥ 4 Carboplatin High (AUC ≥ 4) AUC ≥ 4 Palonosetron 0.25 mg IV, aprepitant 150 mg IV, dexamethasone 10–20 mg IV Ondansetron, prochlorperazine, ± olanzapine Oxaliplatin Moderate 85 or 130 mg/m 2 Palonosetron 0.25 mg IV, dexamethasone 10–20 mg IV Ondansetron, prochlorperazine Open in a new tab 2.2. Blood Samples and Patient‐Reported Outcomes Each patient donated a tube of blood (12 mg K 2 EDTA) the day of their first treatment. Blood samples were collected prior to antiemetics and chemotherapy, at the time medically necessary samples were obtained to determine eligibility for chemotherapy and were subsequently transferred for research analysis. Blood samples were analyzed within 1–2 h of collection using the glutathione recycling kit (Mynari Biomedical LLC) and mixed with the bioactive probe HEDS which is converted into mercaptoethanol (ME) through the enzymatic activities of the glutathione recycling pathway [ 20 , 21 , 23 ]. The amount of ME produced was determined by adding Ellman's reagent to clarified reaction mixtures and measured at 412 nm in a spectrophotometer. Glutathione recycling (GSHrc) capacity was calculated by converting read‐outs to ME per millions of red cells [ 20 , 21 ]. GSHrc results obtained from chemotherapy‐naïve patients were used to predict the occurrence of delayed nausea during any of the subsequent planned treatment cycles. Patients self‐reported nausea symptom and severity using the Rotterdam Symptom Checklist (RSCL) [ 24 ] directly prior to each cycle. Questionnaires were distributed and collected by a study coordinator, who was available to address study‐related questions. Delayed nausea was defined as symptoms occurring 24–120 h following chemotherapy. All patients included in the analysis were prospectively followed for the full duration of their planned treatment course. No changes to antiemetic regimens were made prior to the first report of delayed nausea. Nausea severity was dichotomized into two groups, none or mild and moderate or severe. Mild nausea corresponds to the Common Terminology Criteria for Adverse Events (CTCAE) grade 1, for which no intervention was required. Moderate and severe nausea corresponded to CTCAE grades 2 and 3, respectively, which necessitate changes in management or hospitalization. Reports of nausea were compared to those documented in the electronic medical records. 2.3. Statistical Methods Baseline characteristics were compared between groups using two‐sample t ‐tests, chi‐square tests of independence, and univariable logistic regression to display the effect size of each variable. To incorporate more predictors than the original binary cut‐point identified using CART, elastic net regression was used to build predictive models and a risk score. Elastic net regression with a logit link function was applied to estimate optimal coefficients while minimizing the risk of overfitting. By combining the penalty terms from both LASSO (L1) and Ridge (L2) regression, elastic net enhances model generalizability and improves predictive performance [ 25 ]. The model incorporating the GSHrc score, age, sex, ethnicity, and treatment type was developed using Elastic Net regression. To assess calibration, the dataset was divided into ten groups based on predicted probabilities. The observed outcome incidence was then plotted against the mean predicted probability for each group. Calibration was further quantified using the expected‐to‐observed ratio (E:O) to detect under‐ or over‐estimation of risk, the calibration slope to assess overfitting or underfitting, and the calibration‐in‐the‐large (CITL) to evaluate systematic over‐ or under‐prediction. Model discrimination was measured using the area under the receiver operating characteristic curve (AUC). In addition, 95% confidence intervals for slope and CITL were calculated using logistic regression with predicted probabilities as the only predictor, and for AUC using Stata's roctab command. For internal validation, both 10‐fold cross‐validation and bootstrap cross‐validation were employed. The 10‐fold cross‐validation assessed out‐of‐sample performance by partitioning the data into ten subsets, iteratively training the model on nine parts and validating on the remaining part, and bootstrap cross‐validation was used to estimate model optimism without partitioning the dataset, improving the robustness of internal validation. Validation metrics included the root mean square error (RMSE) and the AUC to reflect both prediction error and discriminative ability. Once the model was validated, a risk score was constructed using the penalized coefficients from the Elastic Net regression. All continuous variables were standardized by Stata before fitting the Elastic Net model. Penalized coefficients were then transformed back to the original scale for the predictors for model presentation and risk‐score calculation. Each score was converted to a predicted probability using the formula: Predicted Probabilities = 1 (1 + exp. (−Risk Score)). The predicted probabilities from the risk score were then validated to ensure they matched the predictive model using a univariable logistic regression model, with nausea severity as the outcome. The risk score probabilities' discrimination and calibration were assessed by calculating the area under the AUC, slope, CITL, and O:E. These metrics matched the original elastic net model metrics. These probabilities were used to estimate the risk of moderate to severe nausea and to identify the corresponding predicted probability risk thresholds. Based on these thresholds, patients were categorized into low, medium, or high‐risk groups. All analyses were conducted using Stata 17.0 (StataCorp LLC, College Station, TX) and R (R Foundation for Statistical Computing, Vienna, Austria). 3. Results After applying exclusion criteria (see Figure 1 ), the analyzed cohort consisted of 202 patients who had a median age of 66 years (range 25–91 years) and follow‐up information for the planned regimen. Younger patients were significantly more likely to report moderate/severe delayed nausea ( p < 0.0001, Table 2 ) than older patients. Enrolled patients predominantly identified as White (81.7%) with Black patients constituting 16.8% of the cohort and the remaining 1.5% representing all other races. Race did not significantly contribute to delayed nausea ( p = 0.745, Table 2 ). Forty‐six percent of participants were female and no significant difference in delayed nausea severity was observed between the sexes ( p = 0.485, Table 2 ). TABLE 2. Demographics and treatment characteristics. Nausea severity OR None/Mild Mod/Severe p n = 157 n = 45 ChemoTox score 1.76 (1.67) 1.58 (1.41) 0.06 0.57 (0.3, 1.0) Age (mean/SD) 67.3 (10.8) 54.2 (13.3) < 0.0001 0.91 (0.88, 0.94) Gender 0.485 Female 71 (45.2%) 23 (51.1%) 1.3 (0.6, 2.4) Male 86 (54.8%) 22 (48.9%) Ref ECOG PS 0 140 (89.2%) 41 (91.1%) 0.351 1 15 (9.5%) 4 (8.9%) 2 2 (1.3%) 0 (0) Race/Ethnicity Caucasian 128 (81.5%) 37 (82.2%) 0.745 Ref African American 27 (17.2%) 7 (15.6%) 0.95 (0.4, 2.3) a Asian 1 (0.6%) 1 (2.2%) Hispanic 1 (0.6%) 0 (0) Chemotherapy < 0.0001 Carboplatin 70 (44.6%) 5 (11.1%) Ref Cisplatin 19 (12.1%) 8 (17.8%) 5.9 (1.7, 20.1) Oxyplatin 68 (43.3%) 32 (71.1%) 6.6 (2.4, 17.9) Open in a new tab a Caucasian vs. non‐Caucasian race/ethnicity. Approximately half of all patients were treated with an oxaliplatin‐containing regimen and 32% of patients in this treatment group reported moderate/severe nausea. Moderate‐to‐severe nausea was reported by 29.6% and 6.7% of patients receiving cisplatin or carboplatin, respectively. Patients reporting moderate/severe nausea were more likely to have been treated with oxaliplatin than the other two agents (Table 2 ). Results from blood testing revealed that those with moderate/severe nausea had a lower mean GSHrc score (1.58 ± 1.41 (SD) vs. 1.76 ± 1.67 (SD)), though the blood assay as a single variable did not reach a statistically significant p ‐value ( p = 0.06). The Elastic net regression identified GSHrc score, age, female, race (non‐white), carboplatin, and cisplatin as coefficients in the model. The predicted probabilities had an AUC of 80.8% (95% CI 74.1%–87.6%), a slope of 1.09 (95% CI 0.73–1.47), a calibration at large (CITL) of 0.09 (95% CI −0.24–0.61), and an E:O of 1 (Figure 2 ). The k ‐fold cross‐validation AUC was 81.6% (95% CI 70.8%, 85.2%), indicating good discrimination and calibration. Based on the identified coefficients, the following risk score algorithm was created: Risk score = 4.33 − 0.51 GSHrc Score − 0.078 age + 0.45 female − 0.16 non − white race − 0.91 carboplatin + 0.41 cisplatin FIGURE 2. Open in a new tab Predicted symptom probability from the Elastic Net Model compared to observed symptoms for each of the ten groups. Calculated predicted probabilities were used to divide patients into groups with low, medium, or high estimated risk for developing delayed nausea (Table 3 ). Patients with a probability of > 50% have at least a high risk of developing moderate/severe delayed nausea; those with 20%–50% are at medium risk, and < 20% are at low risk. In contrast, patients with a low predicted risk for symptoms will have a risk score of ≥ 60. In our cohort, this group constituted 58.9% of patients. Risk scores ranging between 50 and 59 were associated with an intermediate risk of experiencing delayed nausea among 25.7% of patients. TABLE 3. Predicted probability of delayed nausea with risk categories and sample size. Predicted probability Risk n < 10% Low 73 10%–20% Low 42 20%–30% Medium 30 30%–40% Medium 19 40%–50% Medium 13 50%–60% High 13 60%–70% High 6 70%+ High 8 Open in a new tab To determine the performance of the calculated risk score, predicted nausea was compared to experienced nausea within each risk score group. Overall, we found a good correlation between predicted and observed nausea in all risk score groups. Model calibration for delayed chemotherapy‐induced nausea showed good overall agreement between predicted and observed risks (Figure 3 ). In the low‐risk group, the observed incidence was 11.3% compared to the predicted 8.8%. In the medium‐risk group, the model slightly overpredicted nausea (31.4% predicted vs. 25.8%) observed. Calibration was strongest in the high‐risk group where predicted (61.9%) and observed (64.0%) risk were closely aligned. FIGURE 3. Open in a new tab Observed (%) vs. predicted probability (%) of chemotherapy‐induced moderate/severe nausea. The blue bars represent the % of patients who actually had nausea in that risk group, and the red bars represent the % of patients predicted to have nausea within each risk group. 4. Discussion In this study, we present an improvement of our previously published findings on a blood‐based test designed to predict individual risk of delayed chemotherapy‐induced nausea. Elastic net modeling allowed for the inclusion of more variables and use of predicted probabilities on a continuous scale, resulting in higher accuracy in prediction of delayed chemotherapy‐induced nausea. The performance of the prediction algorithm increased the AUC to 0.809 compared to the AUC of 0.716 for the binary cut‐off analysis. Despite advances in the prevention of chemotherapy‐related side effects, CINV—particularly delayed nausea—remains a distressing and costly complication of cancer therapy [ 4 , 6 , 19 , 26 ]. Accurately predicting which patients will develop delayed CINV remains a major clinical challenge [ 26 , 27 ]. Identifying those at highest risk enables clinicians to apply aggressive prophylactic measures selectively, avoiding overtreatment of low‐risk patients. In addition to the considerable reduction in patients' quality of life [ 19 ], the average national cost for an emergency department visit due to severe delayed nausea has been estimated at up to $15,120 [ 28 , 29 , 30 ]. Consequently, multiple intrinsic (e.g., sex, age, race) and extrinsic (e.g., alcohol use, diet, smoking) risk factors have been investigated for their predictive value in CINV [ 31 , 32 , 33 , 34 ]. In our work, we excluded variables were the patients might be reluctant to provide accurate habits, such as alcohol consumption and dietary preferences. A strong prediction factor for CINV is previous CINV symptoms during previous cycles. Therefore, all patients were chemotherapy naïve at the time enrollment. Oxidative stress generated by chemotherapeutic agents in the upper small intestine has been proposed as a key mechanism underlying nausea and vomiting [ 6 , 15 ]. Based on this rationale, we developed the GSHrc assay, which previously demonstrated promising sensitivity and specificity in identifying at‐risk patients using a CART algorithm [ 20 , 21 ]. In the present study, we enrolled patients with solid tumors expected to benefit clinically from platinum‐based therapy. To minimize sex‐related bias—since sex is a known determinant of nausea risk [ 31 , 32 ]—we focused on non–sex‐specific cancers, yielding a balanced cohort of 202 patients (46% female; mean age, 66 years). Cut‐point analysis performed in Stata 17.0 using the Lui option [ 35 ] replicated our earlier observation: patients who developed moderate or severe delayed nausea had lower pre‐treatment GSHrc scores compared to those with no or mild symptoms (1.58 vs. 1.76, p = 0.06). Although not statistically significant, this trend supports the GSHrc score as a potential predictor of delayed CINV. Next, we evaluated established risk factors to determine whether combining intrinsic factors could enhance predictive accuracy. Consistent with prior reports identifying younger age as a risk factor [ 12 , 31 , 36 , 37 ], patients who experienced moderate/severe nausea were significantly younger than those with milder symptoms (54.2 vs. 67.3 years, p < 0.0001). Chemotherapy type was also associated with nausea severity: patients treated with the moderately emetogenic compound oxaliplatin [ 5 , 10 ] accounted for 71.1% of moderate/severe cases ( p < 0.0001). Using a validated model, we generated a composite risk score by applying coefficients derived from Elastic Net regression [ 25 ] to their respective variables, as described in the Results section. This risk score was then validated through univariable logistic regression using nausea severity as the outcome. Among cisplatin‐treated patients ( N = 27), 29.6% reported grade 2/3 nausea, representing 17.8% of all moderate/severe cases. As expected for a highly emetogenic chemotherapy (HEC) [ 5 , 10 ], all cisplatin‐treated patients received standard triple antiemetic prophylaxis with a 5‐HT 3 antagonist (palonosetron), an NK 1 antagonist (aprepitant), and dexamethasone. In contrast, carboplatin ( N = 75)—administered at AUC 2 (MEC) in 80% and AUC 5–6 (HEC) in 20% of patients [ 5 , 10 ]—was associated with significantly lower rates of delayed nausea (6.7% overall, constituting 11.1% of grade 2/3 cases). Notably, oxaliplatin‐treated patients ( N = 100) exhibited the highest incidence of delayed nausea (32%), accounting for the majority (71%) of grade 2/3 events ( p < 0.0001). This finding aligns with observations by Hesketh et al. [ 38 ], who reported delayed nausea in 42.9% of oxaliplatin‐treated patients despite excellent acute‐phase CINV control using a two‐drug prophylaxis (5‐HT 3 antagonist plus dexamethasone). Predicted probabilities derived from our model were categorized into low‐, medium‐, and high‐risk groups (Table 3 ) and compared with observed outcomes (Figure 3 ). Patients with risk scores ≥ 60 were correctly classified as low risk, indicating adequate protection from current standard antiemetic regimens. Scores of 45–59 corresponded to a 25%–50% risk, consistent with previously reported rates of nausea among patients receiving MEC or HEC therapy despite guideline‐concordant prophylaxis [ 3 , 39 , 40 , 41 ]. Clinical trials suggest that these patients benefit from adding an additional antiemetic agent during subsequent treatment cycles [ 5 , 40 ]. In contrast, patients with risk scores ≤ 40 exhibited a predicted 70% and observed > 80% risk of developing delayed nausea, indicating the need for more intensive antiemetic prophylaxis. Our findings, consistent with prior reports [ 38 ], underscore that oxaliplatin‐treated patients contribute disproportionately to delayed nausea incidence when treated with the currently recommended two‐agent regimen. Current prophylaxis guidelines for CINV are largely determined by the emetogenicity of the chemotherapy regimen [ 10 ]. Although major oncology organizations [ 9 , 10 , 42 , 43 ] agree on best practices for both acute and delayed CINV prevention, nausea—more frequent and impactful than vomiting—remains insufficiently controlled. This challenge is compounded by the subjective nature of nausea assessment, in contrast to the objective measurement of vomiting frequency and duration. Incorporating objective biomarkers such as GSHrc, alongside intrinsic factors, may enable a more personalized and precise risk prediction for delayed nausea, reducing bias from patient‐reported measures. Given the differing biological half‐lives of platinum agents, glutathione recycling may play a significant role in patient susceptibility. Carboplatin and cisplatin exhibit half‐lives of 1–2 days, whereas oxaliplatin persists for approximately 5 days. Thus, patients with slower GSHrc rates receiving oxaliplatin may experience prolonged oxidative stress, leading to sustained enterochromaffin cell injury and neurotransmitter release. Such patients may benefit from extended or intensified antiemetic coverage after infusion. Our future goal for the project is to determine the effect of additional antiemetics and the effects on severity and duration of symptoms. Furthermore, the GSHrc test could promote adherence to prescribed oral antiemetics by providing an objective rationale for their continued use. 4.1. Limitations This study has several limitations. First, it was conducted at a single tertiary teaching hospital located in a suburb of Philadelphia, PA, which may limit the generalizability of the findings to broader populations. Second, the outcome measures were based on patient self‐reports. Although we compared patient‐reported outcomes with documentation of nausea symptoms in medical records and adjustments to antiemetic regimens, recall and reporting bias may still have influenced the results due to memory limitations and the time constraints of clinical documentation. Additionally, while all patients received antiemetic therapy consistent with the standard of care—based on the agent with the highest emetogenic potential—and no modifications were made to regimens prior to reported nausea symptoms (as described in the Methods section), we cannot be certain about the extent to which patients used their prescribed PRN antiemetics. Patients who were predicted as medium/high risk of developing delayed CINV but never reported symptoms could have been compliant with their antiemetics and therefore able to prevent any experience of CINV. Third, no external validation was performed to assess the generalizability of the findings to other populations or clinical settings. Although statistical evaluation of the risk score suggests that the findings may be reproducible in similar patient cohorts, a dedicated validation study is needed to confirm these results and reduce potential sources of bias. Future prospective studies are recommended to include a more diverse racial distribution and collect more detailed treatment data, particularly regarding the use of rescue medications. 4.2. Implications of Findings Delayed chemotherapy‐induced nausea remains a major unmet need in supportive oncology care, frequently persisting despite guideline‐recommended antiemetic prophylaxis. Existing risk models, based primarily on demographic and treatment‐related factors, offer limited predictive accuracy. This study introduces a blood‐based glutathione redox cycling assay as an objective biomarker to identify patients at high risk for delayed nausea following platinum‐based chemotherapy. When integrated with clinical variables, GSHrc produced a validated risk score that stratified patients into low‐, intermediate‐, and high‐risk groups, with clear implications for individualized prophylaxis. High‐risk patients—regardless of the emetogenicity of the chemotherapeutic regimen—showed poor symptom control under standard antiemetic protocols and may benefit from intensified preventive therapy. The addition of an NK1 antagonist as a third agent in the antiemetic mixture prior to chemotherapy infusion or a transdermal patch for long‐term delivery of a 5‐HT3 antagonist could be effective approaches to eliminate delayed nausea among the high‐risk patients. Future prospective studies using these agents among high‐risk patients will be needed to ensure efficacy. Incorporating redox‐based biomarkers into CINV risk assessment may enable precision‐guided prevention, enhance patient quality of life, and reduce healthcare utilization associated with uncontrolled nausea. 5. Conclusions Integrating individual patient biology into a multifactorial predictive algorithm enhances the accuracy of identifying those at risk for delayed chemotherapy‐induced nausea. This personalized approach enables clinicians to proactively deliver intensified antiemetic therapy to high‐risk patients, thereby reducing delayed nausea, improving quality of life, and supporting adherence to potentially life‐saving chemotherapy regimens. Author Contributions Study conception and design: P.B.G. and U.M.W.; Methodology and coordination of data collection: A.P.M., K.J.F., Z.A.A., A.G., A.L.S., E.L.Z., D.C.S., and U.M.W.; Data Analysis and original draft: A.P.M., S.D.K., and U.M.W.; Reviewing and editing: A.P.M., S.D.K., K.J.F., E.L.Z., Z.A.A., A.G., A.L.S., D.C.S., P.B.G., and U.M.W. Funding This work has been supported by the Sharpe‐Strumia Research Foundation of Bryn Mawr Hospital, the W. W. Smith Foundation, the Deborah J McKenna Foundation and the Lankenau Institute for Medical Research. Ethics Statement This study was conducted in accordance with the principles of the Declaration of Helsinki . Ethical approval was obtained and maintained throughout the study period from the Lankenau Institutional Review Board (Reference ID: F/N‐R12‐3146L). Written informed consent was obtained from all participants prior to enrollment. Participants were informed that all collected data and test results would remain confidential and used solely for research purposes. Personal identifiers were replaced with study ID numbers on all questionnaires and biological samples. Consent Written informed consent was freely given prior to first treatment visit. Signed consent forms with personal information were securely stored. Conflicts of Interest U.M.W. is the holder of a patent pertaining to the GSHrc assay. All other authors declare no competing financial interests. Supporting information Supplement S1: Modified Rotterdam Symptom Check List CAM4-15-e71824-s001.docx (201.8KB, docx) Acknowledgements This work has been supported by the Sharpe‐Strumia Research Foundation of Bryn Mawr Hospital, the W. W. Smith Foundation, the Deborah J McKenna Foundation and the Lankenau Institute for Medical Research. Data Availability Statement De‐identified data included in this study are available from the corresponding author upon reasonable request. References 1. Cooper S. and Georgiou V., “The Impact of Cytotoxic Chemotherapy—Perspectives From Patients, Specialists and Nurses,” European Journal of Cancer 28A, no. Suppl 1 (1992): S36–S38. [ Google Scholar ] 2. Feyer P. and Jordan K., “Update and New Trends in Antiemetic Therapy: The Continuing Need for Novel Therapies,” Annals of Oncology 22, no. 1 (2011): 30–38. [ DOI ] [ PubMed ] [ Google Scholar ] 3. Grunberg S. M., Deuson R. R., Mavros P., et al., “Incidence of Chemotherapy‐Induced Nausea and Emesis After Modern Antiemetics,” Cancer 100, no. 10 (2004): 2261–2268. [ DOI ] [ PubMed ] [ Google Scholar ] 4. Aapro M., “CINV: Still Troubling Patients After All These Years,” Support Care Cancer 26, no. Suppl 1 (2018): 5–9. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 5. Gupta K., Walton R., and Kataria S. P., “Chemotherapy‐Induced Nausea and Vomiting: Pathogenesis, Recommendations, and New Trends,” Cancer Treatment and Research Communications 26 (2021): 100278. [ DOI ] [ PubMed ] [ Google Scholar ] 6. Navari R. M. and Aapro M., “Antiemetic Prophylaxis for Chemotherapy‐Induced Nausea and Vomiting,” New England Journal of Medicine 374, no. 14 (2016): 1356–1367. [ DOI ] [ PubMed ] [ Google Scholar ] 7. Bosnjak S. M., Zilic A., Radhakrishnan V., et al., “MASCC Antiemetic Consensus Recommendations: Resource‐Limited Settings,” Support Care Cancer 33, no. 3 (2025): 181. [ DOI ] [ PubMed ] [ Google Scholar ] 8. Herrstedt J., Clark‐Snow R., Ruhlmann C. H., Jordan K., and Scotté F., “MASCC/ESMO Antiemetic Guidelines: Introduction to the 2023 Guidelines Update,” Support Care Cancer 32, no. 1 (2023): 57. [ DOI ] [ PubMed ] [ Google Scholar ] 9. Hesketh P. J., Kris M. G., Basch E., et al., “Antiemetics: ASCO Guideline Update,” Journal of Clinical Oncology 38, no. 24 (2020): 2782–2797. [ DOI ] [ PubMed ] [ Google Scholar ] 10. Berger M. J., Ettinger D. S., Aston J., et al., “NCCN Guidelines Insights: Antiemesis, Version 2.2017,” Journal of the National Comprehensive Cancer Network 15, no. 7 (2017): 883–893. [ DOI ] [ PubMed ] [ Google Scholar ] 11. Badarudin N. S., Mohamed Shah N., Mohd Kassim K. N. B., et al., “A Retrospective Study on Chemotherapy‐Induced Nausea and Vomiting in Highly/Moderately Emetogenic Chemotherapy: Incidence and Prescribing Practice,” Support Care Cancer 30, no. 6 (2022): 5339–5349. [ DOI ] [ PubMed ] [ Google Scholar ] 12. Dranitsaris G., Molassiotis A., Clemons M., et al., “The Development of a Prediction Tool to Identify Cancer Patients at High Risk for Chemotherapy‐Induced Nausea and Vomiting,” Annals of Oncology 28, no. 6 (2017): 1260–1267. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Hashimoto Y., Maeda K., Shimomura O., et al., “Development of the Serotonin Release Assay With Enterochromaffin Cell‐Rich Human Intestinal Organoids for the Risk Evaluation of Drug‐Induced Emesis,” Toxicological Sciences 206 (2024): 299–312. [ Google Scholar ] 14. Heitz P., Polak J. M., Timson C. M., and Pearse A. G. E., “Enterochromaffin Cells as the Endocrine Source of Gastrointestinal Substance P,” Histochemistry 49, no. 4 (1976): 343–347. [ DOI ] [ PubMed ] [ Google Scholar ] 15. Hesketh P. J., “Chemotherapy‐Induced Nausea and Vomiting,” New England Journal of Medicine 358, no. 23 (2008): 2482–2494. [ DOI ] [ PubMed ] [ Google Scholar ] 16. Kolli V. K., Abraham P., and Isaac B., “Alteration in Antioxidant Defense Mechanisms in the Small Intestines of Methotrexate Treated Rat May Contribute to Its Gastrointestinal Toxicity,” Cancer Therapy 6 (2008): 17–26. [ Google Scholar ] 17. Raftos J. E., Whillier S., and Kuchel P. W., “Glutathione Synthesis and Turnover in the Human Erythrocyte: Alignment of a Model Based on Detailed Enzyme Kinetics With Experimental Data,” Journal of Biological Chemistry 285, no. 31 (2010): 23557–23567. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 18. Vaskova J., Kočan L., Vaško L., and Perjési P., “Glutathione‐Related Enzymes and Proteins: A Review,” Molecules 28, no. 3 (2023): 1447. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Rapoport B. L., “Delayed Chemotherapy‐Induced Nausea and Vomiting: Pathogenesis, Incidence, and Current Management,” Frontiers in Pharmacology 8 (2017): 19. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 20. Kutner T., Kunkel E., Wang Y., et al., “Preliminary Evaluation of a Predictive Blood Assay to Identify Patients at High Risk of Chemotherapy‐Induced Nausea,” Support Care Cancer 25, no. 2 (2017): 581–587. [ DOI ] [ PubMed ] [ Google Scholar ] 21. McCourt D. D., Parikh K., Brady A. L., et al., “The Quest for Reliable Prediction of Chemotherapy‐Induced Delayed Nausea Among Breast Cancer Patients,” Journal of Unexplored Medical Data 4 (2019): 6. [ Google Scholar ] 22. Waqar S. N., Mann J., Baggstrom M. Q., et al., “Delayed Nausea and Vomiting From Carboplatin Doublet Chemotherapy,” Acta Oncologica 55, no. 6 (2016): 700–704. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 23. Li J., Ward K. M., Zhang D., et al., “A Bioactive Probe of the Oxidative Pentose Phosphate Cycle: Novel Strategy to Reverse Radioresistance in Glucose Deprived Human Colon Cancer Cells,” Toxicology In Vitro 27, no. 1 (2013): 367–377. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 24. de Haes J. C., van Knippenberg F. C., and Neijt J. P., “Measuring Psychological and Physical Distress in Cancer Patients: Structure and Application of the Rotterdam Symptom Checklist,” British Journal of Cancer 62, no. 6 (1990): 1034–1038. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 25. Waldmann P., Mészáros G., Gredler B., Fuerst C., and Sölkner J., “Evaluation of the Lasso and the Elastic Net in Genome‐Wide Association Studies,” Frontiers in Genetics 4 (2013): 270. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 26. Fernandez‐Ortega P., Caloto M. T., Chirveches E., et al., “Chemotherapy‐Induced Nausea and Vomiting in Clinical Practice: Impact on Patients' Quality of Life,” Supportive Care in Cancer 20, no. 12 (2012): 3141–3148. [ DOI ] [ PubMed ] [ Google Scholar ] 27. Foubert J. and Vaessen G., “Nausea: The Neglected Symptom?,” European Journal of Oncology Nursing 9, no. 1 (2005): 21–32. [ DOI ] [ PubMed ] [ Google Scholar ] 28. Nelson W. W., Vaidya V., Scott J. A., et al., “Antiemetic Use and Chemotherapy Induced Nausea and Vomiting Related Hospitalization Costs After Highly or Moderately Emetogenic Chemotherapy,” Future Oncology 19, no. 1 (2023): 29–36. [ DOI ] [ PubMed ] [ Google Scholar ] 29. Roeland E., Ma J., Binder G., et al., “Hospitalization Costs for Nausea and Vomiting: A Savings Opportunity,” Journal of Clinical Oncology 35 (2017): 155. [ Google Scholar ] 30. Sommariva S., Pongiglione B., and Tarricone R., “Impact of Chemotherapy‐Induced Nausea and Vomiting on Health‐Related Quality of Life and Resource Utilization: A Systematic Review,” Critical Reviews in Oncology/Hematology 99 (2016): 13–36. [ DOI ] [ PubMed ] [ Google Scholar ] 31. Mosa A. S. M., Hossain A. M., Lavoie B. J., and Yoo I., “Patient‐Related Risk Factors for Chemotherapy‐Induced Nausea and Vomiting: A Systematic Review,” Frontiers in Pharmacology 11 (2020): 329. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 32. Albanell‐Fernandez M., Rodríguez Mues M. C., Figueras C., et al., “Evaluation of Chemotherapy‐Induced Nausea and Vomiting in Low, Moderate, and Highly Emetogenic Schemes Between Sexes,” Supportive Care in Cancer 33, no. 4 (2025): 261. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 33. Hesketh P. J., Aapro M., Street J. C., and Carides A. D., “Evaluation of Risk Factors Predictive of Nausea and Vomiting With Current Standard‐of‐Care Antiemetic Treatment: Analysis of Two Phase III Trials of Aprepitant in Patients Receiving Cisplatin‐Based Chemotherapy,” Support Care Cancer 18, no. 9 (2010): 1171–1177. [ DOI ] [ PubMed ] [ Google Scholar ] 34. Higa G. M., Auber M. L., and Hobbs G., “Identification of a Novel Marker Associated With Risk for Delayed Chemotherapy‐Induced Vomiting,” Support Care Cancer 20, no. 11 (2012): 2803–2809. [ DOI ] [ PubMed ] [ Google Scholar ] 35. Liu X., “Classification Accuracy and Cut Point Selection,” Statistics in Medicine 31, no. 23 (2012): 2676–2686. [ DOI ] [ PubMed ] [ Google Scholar ] 36. Herrstedt J., Lindberg S., and Petersen P. C., “Prevention of Chemotherapy‐Induced Nausea and Vomiting in the Older Patient: Optimizing Outcomes,” Drugs & Aging 39, no. 1 (2022): 1–21. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 37. Suzuki A., Kobayashi R., Fujii H., et al., “Control of Nausea and Vomiting in Patients With Colorectal Cancer Receiving Chemotherapy With Moderate Emetic Risk,” Anticancer Research 36, no. 12 (2016): 6527–6533. [ DOI ] [ PubMed ] [ Google Scholar ] 38. Hesketh P. J., Sanz‐Altamira P., Bushey J., and Hesketh A. M., “Prospective Evaluation of the Incidence of Delayed Nausea and Vomiting in Patients With Colorectal Cancer Receiving Oxaliplatin‐Based Chemotherapy,” Support Care Cancer 20, no. 5 (2012): 1043–1047. [ DOI ] [ PubMed ] [ Google Scholar ] 39. Jones J. M., Qin R., Bardia A., Linquist B., Wolf S., and Loprinzi C. L., “Antiemetics for Chemotherapy‐Induced Nausea and Vomiting Occurring Despite Prophylactic Antiemetic Therapy,” Journal of Palliative Medicine 14, no. 7 (2011): 810–814. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 40. Navari R. M., “Treatment of Breakthrough and Refractory Chemotherapy‐Induced Nausea and Vomiting,” BioMed Research International 2015 (2015): 595894. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 41. Escobar Y., Cajaraville G., Virizuela J. A., et al., “Incidence of Chemotherapy‐Induced Nausea and Vomiting With Moderately Emetogenic Chemotherapy: ADVICE (Actual Data of Vomiting Incidence by Chemotherapy Evaluation) Study,” Support Care Cancer 23, no. 9 (2015): 2833–2840. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 42. Hesketh P. J., Kris M. G., Basch E., et al., “Antiemetics: American Society of Clinical Oncology Clinical Practice Guideline Update,” Journal of Clinical Oncology 35, no. 28 (2017): 3240–3261. [ DOI ] [ PubMed ] [ Google Scholar ] 43. Roila F., Warr D., Hesketh P. J., et al., “2016 Updated MASCC/ESMO Consensus Recommendations: Prevention of Nausea and Vomiting Following Moderately Emetogenic Chemotherapy,” Supportive Care in Cancer 25, no. 1 (2017): 289–294. [ DOI ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Supplement S1: Modified Rotterdam Symptom Check List CAM4-15-e71824-s001.docx (201.8KB, docx) Data Availability Statement De‐identified data included in this study are available from the corresponding author upon reasonable request. Articles from Cancer Medicine are provided here courtesy of Wiley ACTIONS View on publisher site PDF (648.3 KB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• Comparison of Antiemetic Drugs in Preventing Delayed Nausea After Chemotherapy in Patients With Cancer#1138083
• Predictive Factors for Postoperative Pain, Nausea and Vomiting in Oncological Patients#1172205
• A Multi-Disciplinary Approach to the Management of Chemotherapy-Induced Nausea and Vomiting in Breast Cancer.#367935
• Prevention of Chemotherapy Induced Nausea and Vomiting in Breast Cancer Patients.#847831
• Effect of Aprepitant on Cyclophosphamide Pharmacokinetics in Patients With Breast Cancer#992398