Sage Journals: Discover world-class research

Abstract

Introduction

The Cancer and Aging Research Group (CARG) model predicts chemotherapy-related toxicities in older patients; however, its applicability has not been validated in Taiwanese patients. This study aims to validate the CARG model in older Taiwanese patients with solid tumors.

Methods

Patients (N = 258) aged ≥65 years with solid tumors from a single medical center, slated for first-line chemotherapy, were recruited between 2018 and 2021, with follow-up until December 31, 2022. Patients were categorized into low- (N = 85), medium- (N = 117), and high- (N = 56) risk based on CARG. Validation of CARG involved receiver operating characteristic (ROC) curves. Individual CARG variables were analyzed using univariate analysis for their impact on toxicities and survival.

Results

Toxicities of grades ≥3 were 38.8%, 44.4%, and 67.9% (P = .001) in the three ascending risk groups, and there were significant differences in both hematological (P = .002) and non-hematological (P < .001) toxicities. ROC was 0.631 (95% CI: 0.562-0.700), indicating satisfactory discrimination. One-year overall survival rates were 88.7%, 79.7%, and 63.8%, respectively, in ascending-risk groups, with high-risk groups showing decreased survival (P = .002). In the multivariate analysis, decreased hemoglobin, history of falls, and inability to walk one block remained significantly associated with toxicity. For overall survival, the inability to take medications was the only independent predictor.

Conclusion

This prognostic study validated the CARG model in a heterogeneous solid tumor cohort in Taiwan. In addition to predicting both hematological and non-hematological toxicities, CARG could offer insights into patient survival among older individuals with cancer.

Keywords

adverse event frailty model validation prognosis

Introduction

Advancements in medicine have led to increased life expectancy, resulting in a growing annual diagnosis of cancer among older patients. In 2022, individuals aged ≥65 years constituted over 9.8% of the global population.¹ By 2030, approximately 70% of cancer diagnoses are projected in individuals aged ≥65 years.¹ Despite the challenges associated with advanced age, effective treatment to enhance meaningful survival should not be discouraged.

Although chemotherapy remains integral to cancer treatment, its efficacy is associated with a substantial incidence of treatment-related adverse events. Older adults, due to poor physical reserve, concomitant comorbidities, and a lack of social support, may become particularly susceptible to chemotherapy-related toxicity.² Furthermore, measures to reduce toxicity, such as prescribing reduced dosages, may sometimes compromise treatment efficacy.³ Therefore, administering standard chemotherapy dosages while minimizing the risk of treatment-related toxicity is crucial.

Geriatric assessment (GA), incorporating dimensions such as physical activity, nutrition, mood, cognition, and social function, is widely employed to identify vulnerable older patients undergoing cancer treatment.⁴ Indeed, European Society of Medical Oncology (ESMO) has recommended the implementation of GA and management (GAM) in the care of older cancer patients.⁵ The Cancer and Aging Research Group (CARG) previously constructed a model predicting chemotherapy toxicity in older adults with cancer.⁶ The CARG model consisted of 11 parameters based on patient characteristics, cancer type, number of chemotherapy drugs, chemotherapeutic dosing, laboratory values, and a GA component. Notably, the CARG demonstrated improved toxicity prediction over the Karnofsky Performance Scale, a finding further validated by the same study group.⁷ In a recent prospective study, the CARG was compared to G8 and VES-13 questionnaires, and was found to be more practical in daily use than G8.⁸

In the original CARG study, all patients aged ≥65 years with solid tumors scheduled to receive a new chemotherapy regimen were eligible for inclusion. However, Asians constituted less than 5% of the original cohort.⁶ This is the first study to the authors’ knowledge to validate the CARG model in a population of older Taiwanese patients with solid tumors.

Materials and Methods

Patients

A total of 258 consecutive patients diagnosed with solid tumors were enrolled between 2018 and 2021 at a medical center in Taiwan. The eligibility criteria included age ≥65 years, a scheduled first-line chemotherapy treatment, the ability to complete GA measures, and informed consent. This study was approved by the Linkou Chang Gung Memorial Hospital institutional review board (201600916B0).

The CARG Model

The original CARG model comprised 11 variables: age, tumor site, number of chemotherapy drugs, dosing, laboratory values, and GA questions.⁶ Patients were categorized into three groups based on their total risk score: low- (0-5 points), medium- (6-9 points), and high- (10-19 points) risk, aiming to distribute patients into tertiles.

The GA component included in the CARG model consists of specific questions assessing patient functionality and overall health status. These questions evaluate several key domains: (1) Functional status: Patient-reported ability to perform activities of daily living (ADL), instrumental activities of daily living (IADL), Karnofsky performance status (KPS), incidence of falls, and ability to walk one block; (2) Comorbidities: The total number of medical conditions requiring ongoing treatment or monitoring; (3) Nutritional status: Recent history of unintentional weight loss; (4) Cognitive function: Assessed by the Blessed Orientation-Memory-Concentration test; (5)Psychological state: Evaluated using the Hospital Anxiety and Depression Scale (HADS); and (6) Social support and activity: Patient-reported ability to take medications independently and maintain social activities.

Data Collection

The study procedures mirrored those employed in the original CARG study. All variables in the model were collected prior to planned chemotherapy. Chemotherapy-related toxicities were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (version 3.0) and documented through a chart review. All included patients were followed up until death or December 31, 2022. The same physicians screened the patients as well as provided care for cancer. The reporting of this study conforms to the TRIPOD guidelines for transparent reporting of a multivariable prediction model.⁹

Statistical Analysis

Basic patient demographic data are summarized as proportions (%) for categorical variables and medians with interquartile ranges (IQR) for continuous variables. Differences in patient categorical variables between our study cohort and the original CARG cohort were compared using chi-square tests or Fisher’s exact tests if any cell had fewer than five observations.

Observed grades 3-5 toxicity rates among the three predefined groups were compared using a chi-squared test of proportions. The performance of the model was evaluated using the area under the curve (AUC) of the receiver operating characteristic curves. Additionally, the overall survival (OS) of all patients was compared using Kaplan-Meier curves. Log-rank tests were used to determine the differences between survival curves.

All 11 variables from the CARG model were evaluated using univariate logistic or Cox regression analyses to determine the influence of each variable on chemotherapy-related toxicities and OS. All variables in the univariate analysis with P-values of < .05 were further analyzed and included in the multivariate analysis. All statistical analyses were performed using SPSS software (version 23.0; IBM Corp., Armonk, NY, USA). Statistical significance was set at P < .05.

Results

Patient Characteristics

Table 1 presents a comparison of demographic and clinical variables between the original CARG cohort and ours. In this study’s cohort, mean patient age was 70 years, with 131 males (51%). All 258 patients received chemotherapy as first-line treatment, with 223 (86%) and 196 (76%) receiving standard-dose and polychemotherapy regimens, respectively.

Table 1.

Patient Characteristics.

Characteristics	CARG’s cohort (n = 500)	This cohort (n = 258)	P
Age, years			.18
65-69	175 (35)	114 (44)
70-74	127 (25)	81 (31)
75-79	105 (21)	39 (15)
80-84	73 (15)	18 (7)
≥85	20 (4)	6 (2)
Mean (standard deviation)	73 (6)	70 (4)
Sex			.32
Female	281 (56)	127 (49)
Male	219 (44)	131 (51)
Cancer type			<.001
Breast	57 (11)	56 (17)
Lung	143 (29)	19 (6)
Gastrointestinal	135 (27)	154(46)
Gynecologic	87 (17)	2 (1)
Genitourinary	50 (10)	13 (4)
Other	28 (6)	14 (4)
Cancer stage			<.001
I	23 (3)	17 (7)
II	59 (12)	70 (27)
III	111 (22)	98 (38)
IV	307 (61)	73 (28)
Educational level			<.001
Below high school	18 (4)	140 (56)
High school graduate	175 (35)	80 (31)
Associate/bachelor’s degree	202 (40)	36 (14)
Advanced degree	104 (21)	2 (1)
Missing	1 (0)	0
Race			<.001
Asian	26 (5)	337 (100)
Others	474 (95)	0
Treatment
Standard dose			.10
No	120 (24)	35 (14)
Yes	380 (76)	223 (86)
No. of chemotherapy drugs			.34
Monochemotherapy	149 (30)	62 (24)
Polychemotherapy	351 (70)	196 (76)
Line of chemotherapy			<.001
First line	355 (71)	258 (100)
>First line	145 (29)	0
Growth factor use			<.001
No	294(59)	220 (85)
Yes	206 (41)	38 (15)

Several significant differences were observed in patient characteristics between the two cohorts. Our CARG cohort included more patients with gastrointestinal and breast cancers, stage I-III disease, education below the high school level, Asian race, and receiving chemotherapy in a first-line setting.

Table 2 presents the geriatric assessment results and laboratory values. The original CARG cohort exhibited a higher number of comorbid conditions than our cohort. No statistically significant differences were observed in the functional status, nutritional status, cognition, psychological state, or laboratory values.

Table 2.

Geriatric Assessment and Laboratory Values.

Variable	CARG’s cohort (n = 500)	This cohort (n = 258)	P
Functional status
ADL, mean (SD)	68.5 (26)	94 (11)
IADL, mean (SD)	12.9 (2)	7.4 (1)
KPS, mean (SD)	85.6 (14)	85 (15)
Falls, no. (%)			.24
0	407 (82)	2227 (88)
≥1	91 (18)	31 (12)
Mean (SD)	0.3 (0.8)	0.1 (0.9)
Nutritional status
BMI, mean (SD)	26.2 (5)	23.2 (4)
Unintentional weight loss in the last 6 months, no. (%)			.19
≤6%	328 (66)	157 (61)
>6%	170 (34)	101 (39)
mean (SD)	4.7 (6)	5.0 (6)
No. of comorbid condition, no (%)			<.001
0	48 (10)	72 (28)
1	110 (22)	83 (32)
2	121 (24)	55 (21)
≥3	221 (44)	48 (19)
Cognition, no. (%)
Blessed orientation-memory concentration test score ≥11	29 (6)	22 (9)	.59
Psychological state
Hospital anxiety and depression scale, mean (SD)	8.3 (6)	7.3 (8)
Laboratory value
WBC (k/uL), mean (SD)	7.9 (4)	7.8 (6)
Albumin (g/dL), mean (SD)	3.8 (0.5)	3.6 (0.7)
Hemoglobin (<11 g/dL in males, <10 g/dL in females), no. (%)	62 (12)	52 (20)	.12
Creatinine clearance (Jelliffe, ideal weight) < 34 mL/min, no. (%)	44 (9)	38 (15)	.19
Abnormal liver function tests, N=no. (%)	267 (56)	137 (53)	.67

ADL, activities of daily living; IADL, independent activities of daily living; KPS, Karnofsky performance status; SD, standard deviation; BMI, body mass index; WBC, white blood cell.

Chemotherapy Toxicity

Grade 3-5 chemotherapy-related toxicities are presented in Table 3. Overall, 48% of the patients in our cohort experienced grade 3-5 toxicities compared with 53% of the patients in the original CARG cohort. In our cohort, 29% experienced grade 3-5 hematologic toxicity and 34% experienced non-hematologic toxicity. Whereas in the original CARG cohort, 26% of patients experienced grade 3-5 hematologic toxicity and 43% experienced grade 3-5 non-hematologic toxicity. The most prevalent toxicities in our study were neutropenia (19%), non-neutropenic infections (17%), anemia (12%), and leukopenia (9%). In the original CARG cohort, fatigue (16%), neutropenia (11%), leukopenia (10%), and anemia (10%) were the most common symptoms. Grade 5 toxicity occurred in 1% and 2% of patients in our cohort and the original cohort, respectively.

Table 3.

Grade 3-5 Chemotherapy Toxicity.

Toxicity type	CARG’s cohort (n = 500)				This cohort (n = 258)
Toxicity type	Grade 3-5	Grade 3	Grade 4	Grade 5	Grade 3-5	Grade 3	Grade 4	Grade 5
Hematologic and nonhematologic	265 (53)	197 (39)	58 (12)	10 (2)	123 (48)	98 (38)	24 (19)	1 (1)
Hematologic	131 (26)	90(18)	39 (8)	2 (1)	76 (29)	59 (23)	17 (7)	0
ANC	57 (11)	40 (8)	17 (3)	0	48 (19)	33 (13)	15 (6)	0
WBC	49 (10)	41 (8)	8 (2)	0	23 (9)	18 (7)	5 (2)	0
Hemoglobin	48 (10)	45 (9)	3 (1)	0	31 (12)	30 (12)	1 (1)	0
Platelets	25 (5)	14 (3)	11 (1)	0	12 (5)	9 (4)	3 (1)	0
Infection with abnormal ANC	10 (2)	7 (1)	1 (0)	2 (1)	11 (4)	9 (4)	2 (1)	0
Nonhematologic	217 (43)	184 (37)	25 (5)	8 (2)	88 (34)	72 (28)	15 (6)	1 (1)
Fatigue	81 (16)	79 (16)	2 (1)	0	10 (4)	10 (4)	0	0
Infection with normal ANC	48 (10)	40 (8)	5 (1)	3 (1)	43 (17)	32 (12)	10 (4)	1 (1)
Dehydration	43 (9)	41 (8)	2 (1)	0	5 (2)	4 (2)	1 (1)	0
Thrombosis/embolism	22 (4)	17 (3)	4 (1)	1 (1)	1 (0)	1 (0)	0	0
Hyponatremia	22 (4)	22 (4)	0	0	19 (7)	17 (7)	2 (1)	0
Diarrhea	22 (4)	19 (4)	3 (1)	0	5 (2)	5 (2)	0	0
Hypokalemia	15 (3)	15 (3)	0	0	16 (6)	13 (5)	3 (1)	0
Dyspnea	13 (3)	5 (1)	7 (2)	1 (1)	2 (1)	2 (1)	0	0
Syncope	13 (3)	13 (3)	0	0	2 (1)	2 (1)	0	0
Neuropathy	13 (3)	13 (3)	0	0	0	0	0	0
Nausea	12 (2)	12 (2)	0	0	3 (1)	3 (1)	0	0

ANC, absolute neutrophil count; WBC, white blood cell.

Model Validation

Under CARG risk scoring, the patients were stratified into low-(N = 85, 32.9%), medium-(N = 117, 45.3%), and high-risk (N = 56, 21.7%) groups. The risks of grade 3-5 toxicity were 38.8%, 44.4%, and 67.9% in patients with low-, medium-, and high-risk groups, respectively (P = .001; Figure 1A). The AUC for the mCARG model was 0.631 (95% confidence interval: 0.562-0.700; Supplemental Figure 1).

Figure 1.

The Relationship Between Risk and Grade 3 or Higher Chemotherapy-Related Toxicities. (A) Incidence of These Toxicities Were Significantly Increased in Higher Risk Groups in Both the Original CARG Cohort (P < .001) and Our Patients (P = .001). (B) Both Hematological (P = .002) and Non-hematological (P < .001) Toxicities had Significantly Increased Incidence by CARG Risk Groups.

When examining hematological toxicities, our patients showed significantly increased toxicities in ascending risk groups (24.7%, 23.9%, and 51.8%, P = .002; Figure 1B). Similarly, in terms of non-hematological toxicities, patients also demonstrated significantly increased toxicities in the ascending risk groups (24.7%, 32.5%, and 57.1%, P < .001; Figure 1B).

With a median follow-up time of 33 months (range 2-47 months), 67 patients (26.0%) had died by the end of the study. The overall mortality rates were 14.1%, 27.4%, and 41.1% for patients in the low-, medium-, and high-risk groups, respectively (P = .002; Figure 2).

Figure 2.

Survival of Older Patients Undergoing Chemotherapy After 12, 24, and 36 Months. Patients of Higher Risk Groups had Decreased Survival Compared to the Low-Risk Group.

The hazard ratios were 1.97 (95% CI 1.01-3.82, P = .05) and 3.34 (95% CI 1.66-6.72, P = .001) for medium- and high-risk patients, respectively, compared with low-risk patients.

Univariate and Multivariate Analysis of Toxicity and OS

All 11 variables from the original CARG model were calculated using univariate analyses of both toxicity and overall survival in our cohort (Table 4). Age ≥72 (P = .008), genitourinary (GU) or gastrointestinal (GI) cancer type (P = .005), polychemotherapy (P = .004), decreased hemoglobin (P < .001), experiencing falls (P = .001), and inability to walk one block (P = .01) were significantly associated with toxicity. There were associations to toxicity noted in taking standard dose of chemotherapy, creatinine clearance, hearing, ability to take medications, or decreased social activity. Regarding the impact of survival outcomes, age ≥72 years (P = .048), GU or GI cancer type (P = .02), inability to take medications (P < .001) or walking one block (P = .02) were significant variables associated with poor survival outcome.

Table 4.

Univariate and Multivariate Analysis of Grade 3-5 Chemotherapy Toxicity and Overall Survival.

Variable	Grade 3 to 5 toxicity					Overall survival
Variable	Univariate analysis			Multivariate analysis		Univariate analysis			Multivariate analysis
	b-coefficiency	HR (95%CI)	P	Adjusted HR (95%CI)	P	b-coefficiency	HR (95%CI)	P	Adjusted HR (95%CI)	P
Age ≥ 72 years	0.30	1.35 (1.08-1.67)	.008	1.17 (0.89-1.54)	.26	0.24	1.27 (1.00-1.62)	.048	1.12 (0.88-1.45)	.36
GU or GI cancer type	−0.31	0.73 (0.59-0.91)	.005	0.92 (0.70-1.22)	.58	0.32	1.37 (1.05-1.80)	.022	1.29 (0.98-1.71)	.08
Polychemotherapy	0.41	1.51 (1.14-2.00)	.004	1.36 (0.99-1.86)	.06	−0.04	0.96 (0.73-1.27)	.79
Standard dose	0.11	1.12 (0.85-1.48)	.42			−0.27	0.73 (0.54-1.09)	.15
Hemoglobin	0.34	1.40 (1.16-1.70)	<.001	1.36 (1.07-1.73)	.01	0.10	1.11 (0.91-1.36)	.32
Creatinine clearance	0.14	1.15 (0.92-1.44)	.23			0.17	1.18 (0.95-1.47)	.12
Hearing	0.08	1.08 (0.83-1.40)	.57			0.02	1.02 (0.77-1.36)	.88
Falls	0.31	1.37 (1.09-1.73)	.008	1.48 (1.11-1.98)	.008	0.16	1.17 (0.94-1.45)	.15
Take medication	0.32	1.38 (0.72-2.63)	.34			1.32	3.74 (2.06-6.80)	<.001	2.78 (1.38-5.58)	.004
Walking one block	0.63	1.88 (1.28-2.75)	.01	1.89 (1.07-3.35)	.030	0.42	1.52 (1.08-2.15)	.017	1.20 (0.79-1.82)	.39
Decreased social activity	0.14	1.15 (0.64-2.08)	.64			−0.03	0.97 (0.51-1.86)	.94
Increment in score	0.17	1.19 (1.10-1.28)	<.001			0.12	1.13 (1.05-1.21)	.001

GI, gastrointestinal; GU, genitourinary.

To further explore the independent predictors of toxicity and overall survival, we conducted multivariate analyses including variables that reached statistical significance in univariate models (P < .05). In the multivariate analysis, decreased hemoglobin, history of falls, and inability to walk one block remained significantly associated with toxicity. For overall survival, the inability to take medications was the only independent predictor.

Discussion

This study revealed that chemotherapy toxicity is prevalent in older patients, with nearly half experiencing grade ≥3 adverse events, and 1% succumbing to treatment-related complications. The original CARG study predicted chemotherapy-related toxicity in older patients. Our Taiwanese CARG study similarly demonstrated effective risk discrimination. The model also successfully predicted toxicity by hematological and non-hematological adverse event categories. Additionally, the CARG model successfully predicted overall survival trends in our patients.

Numerous studies have sought to validate the CARG model in diverse populations, often in comparison with other evaluation methods. Suto et al¹⁰ applied CARG to a Japanese population, where the model significantly demonstrated validity in predicting chemotherapy toxicity.¹⁰ Additionally, they observed that Eastern Cooperation Oncology Group (ECOG) performance status did not serve as a reliable predictor.¹⁰ Similarly, in India, a large prospective observational study found CARG to be successful in predicting toxicity, whereas ECOG performance status failed to do so.¹¹ Another prospective study in Germany, involving 104 patients, compared CARG with Geriatric-8 (G8) and concluded that CARG strongly predicted toxicity, whereas G8 did not.¹² In Canada, CARG was validated in a large population of patients with metastatic castration-resistant prostate cancer.¹³ Recently, Pang et al also identified the valid application of CARG in a multiethnic Asian population in Singapore.¹⁴

However, not all studies have corroborated these findings. In a French ELCAPA cohort of patients aged >70 years, Frelaut et al¹⁵ discovered that CARG was a poor predictor of chemotherapy toxicity.¹⁵ Similarly, Chan et al¹⁶ evaluated CARG and G8 in a Hong Kong population¹⁶ and found that both tools were ineffective predictors of toxicities. However, they observed higher CARG scores were associated with unexpected hospitalization.¹⁶ Other factors, such as pain, may be equally crucial in addressing chemotherapy toxicity.¹⁷ These discrepant findings may be explained by the differences in study methodology, patient population, and type of cancer intervention provided.

Similar to the original CARG study, our Taiwanese cohort demonstrated adequate risk discrimination ability. Although the CARG study identified 11 variables significant to toxicity prediction, our study only found 6 of them to be associated with higher toxicities, including age ≥72 years, GU or GI cancer type, polychemotherapy, decreased hemoglobin, falls, and the inability to walk one block. Several factors may have contributed to the disparity in toxicity evaluation between the original CARG cohort and our study. The model, partially relying on patient subjective input, may have been affected by differences in education and cultural variance between the two cohorts, potential masking the ability of the model to predict treatment toxicity in Asian populations. Our cohort primarily possessed educational backgrounds below the high school level, whereas the CARG cohort predominantly held bachelor’s degree or higher. Furthermore, side effects encompass partly subjective parameters such as fatigue, syncope, and neuropathy, serving as vulnerability admissions. Interestingly, fatigue, being the parameter with the greatest subjective influence, exhibited the most prominent distinction between the two groups. This discrepancy may mirror the resilience and cultural attitudes within the population, thereby influencing variations in toxicity.

There were also other observable differences between the original CARG cohort and ours. Notably, our patients had lower stage disease, and were exclusively receiving first-line chemotherapy. In comparison, the original cohort patients had mostly stage IV disease, many of whom were receiving further line treatments. Previous studies have shown that further-line therapies may incur increasing toxicities with diminishing returns on survival.¹⁸ These disparities may have contributed to the numerically lower frequencies of chemotherapy-related toxicities across all risk groups in our study patients.

Interestingly, this study found that the CARG model exhibited significant correlation with OS. However, literature on this phenomenon is currently lacking. Incidentally, a small prospective observational study of non-colorectal gastrointestinal cancer observed that CARG could effectively predict OS.¹⁹ The CARG model incorporated age, sociodemographic factors, tumor and treatment characteristics, laboratory data, and GA variables for toxicity prediction, these same parameters are also applicable for survival prediction in the same population.^20,21 Anemia, kidney failure, fall-related injuries—integral components of CARG—may contribute to higher morbidity and mortality.^22-24 ECOG performance status of ≥2 would also serve as an indicator of poorer survival.²⁵

Several of these parameters serve as indicators of frailty. In Shamliyan’s systematic review of 24 studies, weight loss, fatigue/exhaustion, weakness, low physical activity/slowness, and mobility impairment were identified as key parameters for frailty.²⁶ The review highlighted the prevalence of frailty in geriatric syndromes and its correlation with reduced survival.²⁶ Handforth et al²⁷ found that over half of older patients with cancer experienced frailty, leading to increased treatment complications and all-cause mortality.²⁷ In the CARG model, frailty is represented by factors such as the ability to take medication, walking one block, experiencing falls, and reduced social activity.⁶ In this study, the ability to take medication and walking one block were found to be associated with survival. Further research exploring the relationship between CARG scores and survival is warranted.

Several limitations exist within this study. Patient side effects were assessed with CTCAE version 3 instead of version 4 due to the time period of data collection. The role of dose reduction in regards to toxicity is unclear and its data collection incomplete, although monochemotherapy, polychemotherapy, lines of chemotherapy use are recorded for consideration. Additionally, hematological cancers were not included in this study; however, this inclusion may provide more insights into CARG applicability to a wider range of patients with cancer. Future research may include larger numbers of patients with hematological malignancies for broader generalizability of the CARG.

Our CARG investigation focused on a Taiwanese population with mixed solid tumors. In addition to successful validation across all chemotherapy-related toxicities as well as in hematological and non-hematological toxicity subgroups, CARG demonstrated good survival prediction. Therefore, our study suggests that CARG serves as a reliable tool for assessing the adverse effects of chemotherapy on older adults. To further enhance the broader applicability of CARG in patients with universal cancers, additional cohort studies are warranted.

Supplemental Material

Supplemental Material - Validation of a Chemotherapy Toxicity Prediction Model in Older Adults With Cancer in Taiwan

Supplemental Material for Validation of a Chemotherapy Toxicity Prediction Model in Older Adults With Cancer in Taiwan by Chieh-Ying Chang, Yu-Shin Hung, Ming-Chung Kuo, and Wen-Chi Chou in Cancer Control

Footnotes

Authors’ Note

The abstract of this paper was presented at the ASCO Breakthrough Conference as a poster presentation with interim findings. The poster’s abstract was published in ‘Poster Abstracts’ in JCO 42, 223-223 (2024) with DOI:10.1200/JCO.2024.42.23_suppl.223.

Acknowledgements

The authors gratefully acknowledge the assistance of the patients who participated in this study.

ORCID iD

Chieh-Ying Chang

Ethical Statement

Author Contributions

Conception and design of study: CYC, YSH, MCK, WCC; Acquisition of data: YSH, MCK; Analysis and interpretation of data: CYC, YSH, MCK, WCC; Drafting of the manuscript: CYC, YSH, MCK, WCC.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by research grants from the Chang Gung Memorial Hospital at Linkou, Taoyuan, Taiwan, R.O.C. (CMRPG3L1611 and CORPG3N0151); the Ministry of Health and Welfare, Taiwan, R.O.C. (MOHW112-TDU-B-222-124011); and the National Science and Technology Council, Taiwan, R.O.C. (NSTC 111-2314-B-182A-162 and NSTC 112-2314-B-182A-152).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The datasets used and/or analyzed in the current study are available from the corresponding author upon reasonable request.*

Supplemental Material

Supplemental material for this article is available online.

References

Organisation for Economic Co-Operation and Development . Elderly Population. Paris: OECD.org; 2024.

Fried

Ferrucci

Darer

Williamson

Anderson

. Untangling the concepts of disability, frailty, and comorbidity: implications for improved targeting and care. J Gerontol A Biol Sci Med Sci. 2004;59(3):M255-M263.

Lyman

. Impact of chemotherapy dose intensity on cancer patient outcomes. J Natl Compr Canc Netw. 2009;7(1):99-108.

Disalvo

Moth

Soo

, et al. The effect of comprehensive geriatric assessment on care received, treatment completion, toxicity, cancer-related and geriatric assessment outcomes, and quality of life for older adults receiving systemic anti-cancer treatment: a systematic review. J Geriatr Oncol. 2023;14(8):101585.

Loh

Liposits

Arora

, et al. Adequate assessment yields appropriate care the role of geriatric assessment and management in older adults with cancer: a position paper from the ESMO/SIOG cancer in the elderly working group. ESMO Open. 2024;9(8):103657.

Hurria

Togawa

Mohile

, et al. Predicting chemotherapy toxicity in older adults with cancer: a prospective multicenter study. J Clin Oncol. 2011;29(25):3457-3465.

Hurria

Mohile

Gajra

, et al. Validation of a prediction tool for chemotherapy toxicity in older adults with cancer. J Clin Oncol. 2016;34(20):2366-2371.

Cavdar

Iriagac

Karaboyun

Avci

Seber

. Prospective comparison of the value of CARG, G8, and VES-13 toxicity tools in predicting chemotherapy-related toxicity in older Turkish patients with cancer. J Geriatr Oncol. 2022;13(6):821-827.

Collins

Reitsma

Altman

Moons

KGM

. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement. BMJ. 2015;350:g7594.

10.

Suto

Inui

Okamura

. Validity of the cancer and aging research group predictive tool in older Japanese patients. Cancers (Basel). 2022;14(9):2075.

11.

Ostwal

Ramaswamy

Bhargava

, et al. Cancer Aging Research Group (CARG) score in older adults undergoing curative intent chemotherapy: a prospective cohort study. BMJ Open. 2021;11(6):e047376.

12.

Kotzerke

Moritz

Mantovani

, et al. The performance of three oncogeriatric screening tools - G8, optimised G8 and CARG - in predicting chemotherapy-related toxicity in older patients with cancer. A prospective clinical study. J Geriatr Oncol. 2019;10(6):937-943.

13.

Alibhai

SMH

Breunis

Hansen

, et al. Examining the ability of the Cancer and Aging Research Group tool to predict toxicity in older men receiving chemotherapy or androgen-receptor-targeted therapy for metastatic castration-resistant prostate cancer. Cancer. 2021;127(14):2587-2594.

14.

Pang

Jiali

, et al. Use of the cancer and aging research group predictive model for chemotherapy-related toxic effects in a multiethnic, older adult Asian population. JAMA Netw Open. 2022;5(10):e2237196.

15.

Frelaut

Paillaud

Beinse

, et al. External validity of two scores for predicting the risk of chemotherapy toxicity among older patients with solid tumors: results from the ELCAPA prospective cohort. Oncologist. 2023;28(6):e341-e349.

16.

Chan

Cheung

, et al. The predictive value of G8 and the cancer and aging research group chemotherapy toxicity tool in treatment-related toxicity in older Chinese patients with cancer. J Geriatr Oncol. 2021;12(4):557-562.

17.

Bergerot

Razavi

Philip

, et al. Association between chemotherapy toxicity risk scores and physical symptoms among older Brazilian adults with cancer. J Geriatr Oncol. 2020;11(2):280-283.

18.

Gupta

Eisenhauer

Booth

. The time toxicity of cancer treatment. J Clin Oncol. 2022;40(15):1611-1615.

19.

Nishijima

Deal

Williams

Sanoff

Nyrop

Muss

. Impact of the cancer and aging research group score and treatment intensity on survival and toxicity outcomes in older adults with advanced noncolorectal gastrointestinal cancers. Cancer. 2022;128(10):1929-1936.

20.

Denewet

De Breucker

Luce

Kennes

Higuet

Pepersack

. Comprehensive geriatric assessment and comorbidities predict survival in geriatric oncology. Acta Clin Belg. 2016;71(4):206-213.

21.

van Walree

Scheepers

van Huis-Tanja

, et al. A systematic review on the association of the G8 with geriatric assessment, prognosis and course of treatment in older patients with cancer. J Geriatr Oncol. 2019;10(6):847-858.

22.

Lameire

Flombaum

Moreau

Ronco

. Acute renal failure in cancer patients. Ann Med. 2005;37(1):13-25.

23.

Caro

Salas

Ward

Goss

. Anemia as an independent prognostic factor for survival in patients with cancer. Cancer. 2001;91(12):2214-2221.

24.

Wildes

Dua

Fowler

, et al. Systematic review of falls in older adults with cancer. J Geriatr Oncol. 2015;6(1):70-83.

25.

Dall’Olio

Maggio

Massucci

Mollica

Fragomeno

Ardizzoni

. ECOG performance status ≥2 as a prognostic factor in patients with advanced non small cell lung cancer treated with immune checkpoint inhibitors—a systematic review and meta-analysis of real world data. Lung Cancer. 2020;145:95-104.

26.

Shamliyan

Talley

KMC

Ramakrishnan

Kane

. Association of frailty with survival: a systematic literature review. Ageing Res Rev. 2013;12(2):719-736.

27.

Handforth

Clegg

Young

, et al. The prevalence and outcomes of frailty in older cancer patients: a systematic review. Ann Oncol. 2015;26(6):1091-1101.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.05 MB