Radiologic Sarcopenia and Charlson Comorbidity Index in the Head and Neck Free Flap Population

ABSTRACT

Objective

Determining operative risk in patients undergoing head and neck free flap surgery has traditionally been difficult. Psoas muscle mass, a proven frailty marker in abdominal surgery, has not been evaluated in this population. This study demonstrates that radiologic sarcopenia and comorbidity burden may identify patients at higher risk for non‐home discharge.

Methods

A retrospective review was performed on 308 patients who underwent head and neck free flap surgery and had preoperative abdominal CT scans between 2014 and 2024. Inverse Psoas Muscle Index (PMI) and Charlson Comorbidity Index (CCI) were analyzed to assess the impact of radiologic sarcopenia and comorbidity on discharge disposition.

Results

Higher inverse PMI had increased odds of discharge non‐home (OR 1.37, 95% CI 1.05–1.77, p = 0.019). Non‐home discharge varied across CCI groups (p < 0.05), occurring in 7.4% of patients with mild comorbidity, 39.5% with moderate comorbidity, and 55.7% with high comorbidity. Higher CCI increased odds of discharge non‐home (OR 1.30, 95% CI 1.13–1.49, p < 0.001). Highest PMI quartile and mild comorbidity resulted in a 100% PPV for discharge home, corresponding to a 0% non‐home discharge rate.

Conclusion

Lower psoas muscle mass, reflected by higher inverse PMI, independently predicts non‐home discharge, supporting its role as a frailty marker in head and neck free flap patients. Patients with higher comorbidity burden, as measured by CCI, have higher odds of non‐home discharge. Combining radiologic muscle assessment (inverse PMI) with comorbidity indices demonstrates strong predictive potential for non‐home discharge.

This retrospective review evaluates whether radiologic sarcopenia and comorbidity burden predict non‐home discharge following head and neck free flap reconstruction. Lower psoas muscle mass (higher inverse PMI) and greater comorbidity burden (higher CCI) independently predicted non‐home discharge, suggesting these readily available preoperative measures may improve risk stratification, discharge planning, and patient counseling.

1. Introduction

Determining operative risk in patients undergoing head and neck free flap reconstruction has traditionally been difficult. Many patients present with significant complicated comorbidities that preclude objective risk assessment based on clinical characteristics alone. To improve risk stratification based on patient comorbidities, previous approaches to determine operative risk have adapted objective assessment models originally designed for other contexts. The Charlson Comorbidity Index was developed to classify comorbid conditions and predict mortality in medical patients and has been validated to independently predict short‐ and long‐term mortality in acutely ill hospitalized elderly adults [1, 2]. In head and neck surgery, the index has been linked to increased length of hospital stay and flap survival [3, 4].

To introduce a more standardized method of preoperative risk stratification and to quantify the surgeon's subjective assessment of patient physical status, previous authors have proposed the novel field of analytic morphomics to measure patient frailty [5, 6]. Frailty is a syndrome of decreased reserve and resistance to stressors that causes vulnerability to poor outcomes [7]. To gauge frailty, clinicians measure muscle area and density on standardized cuts from cross‐sectional imaging and compare measurements across cohorts with otherwise similar comorbidities. Core muscle size, defined as measurements taken on the psoas muscle on computed tomography (CT) scan, has been previously used as an objective marker of surgical frailty and, by extension, functional status in patients undergoing elective general and vascular surgery [5, 6, 8]. One study in elderly trauma patients, showed that each 1 cm² increase in psoas area was linked to 20% higher odds of independent discharge [9]. In patients who undergo abdominal imaging as part of their oncologic evaluation or staging, psoas muscle measurements can be obtained retrospectively without additional imaging or patient burden. While morphomic analysis of the temporalis muscle has been used to predict postoperative outcomes in head and neck cancer, analysis of psoas muscle measurements in comparison to other commonly used measures of comorbidity‐based preoperative risk has not been performed [10].

In this study, we examine the utility of core muscle size and the CCI as objective measures of frailty and predictors of discharge disposition in patients undergoing head and neck free flap reconstruction. We hypothesized that higher inverse PMI (reflecting lower psoas muscle mass) and higher CCI scores are associated with increased odds of non‐home discharge.

2. Materials and Methods

The study was approved by the institutional review board of Loma Linda University Health and conducted in accordance with the Declaration of Helsinki. A retrospective chart review was conducted on 308 patients who underwent primary head and neck free flap surgery and had preoperative abdominal CT scans between September 1, 2014, and September 1, 2024. Data collection included demographics (age, sex, race, body mass index [BMI], smoking status, and tracheostomy), medical history, surgical factors, and postoperative outcomes (disposition, mortality). Psoas muscle area was measured bilaterally at L3, and Psoas Muscle Index (PMI) was calculated as total area divided by patient height squared (cm²/m²) [2]. For analysis, the inverse of PMI (1/PMI) was used to examine its association with discharge disposition. CCI was calculated on a 37‐point scale and was grouped by points: mild (0–2), moderate [3, 4, 5], high (≥ 6) comorbidity. The outcome studied was patient disposition, categorized as either being discharged to non‐home (skilled nursing, rehabilitation, hospital, hospice, or death within 30 days of surgery) or home.

2.1. Statistical Analysis

Baseline characteristics were summarized using descriptive statistics, with continuous variables reported as mean ± standard deviation or median (interquartile range) and categorical variables as counts and percentages. Differences in CCI groups and inverse PMI by discharge disposition were assessed using Chi‐square and Mann–Whitney U tests, respectively. Univariate logistic regression analyses were performed to evaluate predictors of discharge disposition, including patient demographics (age, sex, race, BMI, smoking status, and tracheostomy), inverse PMI, and CCI. Multivariate logistic regression analysis was performed to evaluate predictors of discharge disposition, including patient demographics (sex, race, BMI, and smoking status) and inverse PMI; age and tracheostomy were excluded due to their previously reported associations in the literature. Multivariate logistic regression analysis was also performed for discharge disposition in respect to patient demographics (sex, race, BMI, smoking status) and CCI. Analyses were conducted in R version 4.4.1 (2024‐06‐14) via the R editor using the dplyr (v1.1.4), ggplot2 (v3.5.1), jmv (v2.5.6), jmvcore (v2.4.7), and Rcpp (v1.0.13) packages. Supplementary analyses were performed in Jamovi (v2.6.2.0) and Microsoft Excel for Mac (v16.83). Statistical significance was determined at p < 0.05.

3. Results

Patient demographics are summarized in Table 1. Higher inverse PMI had increased odds of non‐home discharge (Figure 1). Non‐home discharge differed significantly across CCI groups (p < 0.05), occurring in 7.4% (23/308) of patients with mild comorbidity, 39.5% (122/308) with moderate comorbidity, and 55.7% (171/308) with high comorbidity (Figure 2). For every 1‐point increase in inverse PMI, there was a 37% increase in the odds of non‐home discharge, which was statistically significant (OR 1.37, 95% CI 1.05–1.77, p = 0.019) (Figure 3). Higher CCI increased odds of non‐home discharge: each 1‐point increase in CCI raised the odds of non‐home discharge by 30% (OR 1.30, 95% CI 1.13–1.49, p < 0.001) (Figure 3). On multivariate analysis, both inverse PMI and CCI remained independent predictors of non‐home discharge, with each 1‐point increase in inverse PMI associated with 34% higher odds (OR 1.34, 95% CI 1.00–1.80, p = 0.049) and each 1‐point increase in CCI associated with 31% higher odds (OR 1.31, 95% CI 1.14–1.51, p < 0.001) (Figure 4). The positive predictive value (PPV) for discharge home varied across PMI quartiles and CCI groups (Table 2). Patients in the highest PMI percentile (75th‐100th) had a PPV of 70.1% for discharge home, compared with 54.3% among those in the lower PMI percentile (0th‐75th). Patients with mild CCI (0–2 points) demonstrated a PPV of 92.6% for home discharge; those with moderate CCI (3–4 points) had a PPV of 60.5%, ansd those with high CCI (≥ 6 points) had a PPV of 44.3%. When both favorable characteristics were present—patients in the highest PMI quartile with mild CCI—the PPV for home discharge reached 100%, corresponding to a 0% non‐home discharge rate.

TABLE 1.

Patient demographics (n = 308).

Variable	Value
Age (years)	64 (57–76)
Female	111 (36.0%)
BMI (kg/m2)	25.7 ± 6.4
Current Smoker	24 (7.8%)
Race
Non‐Hispanic White	204 (66.2%)
Hispanic	67 (21.8%)
African American	14 (4.5%)
Asian	11 (3.6%)
Other or Patient Refused	12 (3.9%)
Disposition Location
Home	179 (58.1%)
Home or Self Care	73 (23.7%)
Home Under Care of OHHC	106 (34.4%)
Non‐home	129 (41.19%)
Skilled nursing facility	77 (25.0%)
Inpatient rehabilitation	27 (8.8%)
Inpatient short‐term general hospital	1 (0.3%)
Inpatient long‐term care hospital	19 (6.2%)
Hospice	2 (0.7%)
Death (within 30 days of surgery)	3 (1.0%)
CCI Score (scale 0–37)	4.91 ± 1.85
PMI (cm2/m2)	2.94 ± 1.12

Note: Results are presented as median (IQR) or mean ± SD for continuous variables, and n (%) for categorical variables.

Abbreviations: CCI, Charlson Comorbidity Index; OHHC, organized home health service; PMI, psoas muscle index.

FIGURE 1.

Non‐home discharge percentage by inverse PMI with trendline.

FIGURE 2.

Non‐home discharge percentage by CCI group.

FIGURE 3.

Univariate logistic regression for patient demographics, inverse PMI, and CCI, on non‐home disposition. Reference level: “Disposition = 1 (home).” *p < 0.05 indicates statistical significance.

FIGURE 4.

Multivariate logistic regression for patient demographics, inverse PMI, and CCI, on non‐home disposition. Reference level: “Disposition = 1 (home).” *p < 0.05 indicates statistical significance.

TABLE 2.

Positive predictive values for discharge home.

Variable	Group	PPV (%)
PMI	0th–75th percentile	54.3
	75th–100th percentile	70.1
CCI	Mild (0–2 points)	92.6
	Moderate (3–4 points)	60.5
	High (≥ 6 points)	44.3
Combined PMI & CCI	Highest PMI percentile (75th–100th) + Mild CCI	100

4. Discussion

Head and neck cancer surgery carries a high risk of potentially life‐altering complications, and accurately predicting individual patient outcomes has historically been difficult. Traditional predictors, including BMI, sex, chronologic age, and comorbidity indices, have shown inconsistent associations with outcomes [11, 12, 13, 14]. The effect of chronologic age on outcomes has been more difficult to quantify, with studies offering conflicting results depending on sample size and patient characteristics. Frailty, often conceptualized as “physiologic age,” may better capture this heterogeneity and has been increasingly recognized as a predictor of surgical outcomes [14, 15, 16, 17, 18, 19].

Assessment of frailty, however, remains challenging due to the lack of a standardized definition or universally agreed‐upon measurement. Various indices, including ASA classification, the CCI, and institutional frailty indices, have been used, but these measures primarily reflect comorbidity burden rather than functional reserve. In contrast, morphomic analysis, particularly measurement of core muscle mass on cross‐sectional imaging, offers an objective and reproducible method to estimate frailty [20, 21].

The psoas muscle has been identified as a target of frailty research due to its critical function in spinal stability, control, and ambulation [22, 23]. Prior studies have demonstrated that decreased psoas muscle area correlates with higher postoperative morbidity and mortality in diverse surgical populations, including liver transplantation and abdominal aortic aneurysm repair [5, 6]. Other morphomic metrics such as low psoas density have been shown to predict poor post‐operative outcomes in cancer patients [24, 25]. However, evidence in head and neck cancer patients remains limited, especially regarding discharge disposition. Although recent studies have investigated cervical metrics and have shown a strong predictive value of skeletal muscle cross‐sectional area at the C3 level for the area at L3, the psoas muscle at L3 remains the most validated site for estimating whole‐body skeletal muscle mass [26]. Applying this well‐established abdominal frailty marker to head and neck free flap patients extends its use to a population that has been underrepresented in the frailty literature and offers a consistent benchmark for comparison with other surgical specialties.

Growing evidence also supports an association between sarcopenia and adverse outcomes specifically in head and neck cancer surgery. In patients undergoing total laryngectomy, radiographically defined sarcopenia has been associated with increased postoperative and wound complications, including pharyngocutaneous fistula [27]. Other studies of head and neck free flap reconstruction have demonstrated associations between reduced skeletal muscle mass and discharge to post‐acute care facilities, as well as increased perioperative transfusion requirements [28]. Higher CCI has also been associated with non‐home discharge and prolonged hospitalization following head and neck free flap reconstruction [29]. Together, these findings suggest that radiographic muscle measurements and comorbidity indices capture distinct but clinically relevant dimensions of preoperative vulnerability.

Our study extends these findings to head and neck free flap surgical patients and demonstrates that higher inverse PMI (reflecting lower psoas muscle mass) and higher CCI scores independently predicted increased odds of non‐home discharge. Importantly, the combination of high PMI and low CCI yielded a 100% positive predictive value for home discharge, corresponding to a 0% non‐home discharge rate, suggesting that integrating both radiologic and comorbidity‐based assessments enhances risk stratification. These findings support the utility of psoas muscle measurement as a complementary tool to traditional comorbidity indices for preoperative planning and counseling in this population.

Many studies have operated under the assumption that psoas muscle measurements are correlated with skeletal muscle mass [30, 31, 32]. However, other studies have questioned whether a single CT slice can accurately represent whole‐body muscle, with whole‐body imaging remaining the gold standard for quantitative sarcopenia assessment [33, 34]. Nevertheless, prior work has shown that abdominal CT slices correlate reasonably well with overall adipose and lean tissue, supporting their practical use in clinical research [35, 36]. Differences in study populations, measurement techniques, and outcomes may account for the variability in reported associations between psoas‐based sarcopenia and postoperative outcomes.

This study has several important limitations. The most frail patients may not have been offered surgery, introducing selection bias and limiting generalizability to the broader head and neck cancer population. Inclusion was restricted to patients with preoperative abdominal CT scans, which may bias the cohort toward sicker patients or those with metastatic disease. Although the CCI captures comorbidity burden, it does not fully reflect functional frailty, physiologic reserve, or surgical complexity, potentially underestimating certain surgical risks. Psoas muscle area represents a radiographic measure of skeletal muscle mass rather than a comprehensive assessment of sarcopenia, which may also incorporate measures of muscle strength and physical performance. Accordingly, radiologic PMI does not fully capture functional status or physiologic reserve. The retrospective design and single institution setting further limit the applicability of these findings. Future prospective studies incorporating broader patient populations and standardized frailty assessments may help validate these findings and improve prediction of postoperative disposition in head and neck free flap surgery.

5. Conclusion

Lower psoas muscle mass, reflected by higher 1/PMI, independently predicts non‐home discharge, supporting its role as a frailty marker in head and neck free flap patients. Patients with higher comorbidity burden, as measured by CCI, have higher odds of non‐home discharge. Combining radiologic psoas muscle assessment with comorbidity indices augments traditional frailty assessment with an objective biomarker and demonstrates strong predictive potential. These findings suggest that incorporating frailty and comorbidity indices into preoperative evaluation may help clinicians anticipate recovery trajectories, optimize discharge planning, and improve education for patients and their families.

Funding

The authors have nothing to report.

Disclosure

Artificial intelligence was not used in the preparation of this manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Vacaru A., Qu R. W., Hauser J., et al., “Radiologic Sarcopenia and Charlson Comorbidity Index in the Head and Neck Free Flap Population,” Laryngoscope Investigative Otolaryngology 11, no. 5 (2026): e70580, 10.1002/lio2.70580.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.