Looking Beyond the Eyeball Test: A Novel Vitality Index to Predict Recovery after Esophagectomy

Abstract

Objectives:

To 1) measure 4 physiologic metrics prior to esophagectomy and use these in an index to predict composite postoperative outcome after esophagectomy, and 2) compare predictive accuracy of this index to that of the Fried Frailty Index and Modified Frailty Index.

Methods:

Grip strength (kg), 30-second chair sit-stands (number), 6-minute walk distance (m), and normalized psoas muscle area (cm²/m) were measured for 77 consenting patients from 1/1/2018 to 4/1/2019. Imbalanced random forest classification estimated probability of a composite postoperative outcome, which included mortality, respiratory complications, anastomotic leak, delirium, length of stay ≥14 days, discharge to nursing facility, and readmission. G-mean error was used to compare predictive accuracy among indexes.

Results:

Median grip strength was 38 kg [25th-75th percentiles, 31–44], number of sit-stands 11 [10–14], psoas muscle area to height ratio 6.9 cm²/m [6.0–8.2], and 6-minute walk distance 407 m [368–451]. There was generally weak correlation between these metrics, with the highest between 30-second sit-stands and 6-minute walk distance (r=0.57). Age, degree of patient-reported exhaustion, and the 4 objective metrics comprised the Esophageal Vitality Index, which had a lower G-mean error of 32% [31–33] than the Fried Frailty Index, 37% [37–38], and the Modified Frailty Index, 48% [47–48].

Conclusion:

The Esophageal Vitality Index, an objective, simple assessment consisting of grip strength, 30-second chair sit-stands, 6-minute walk, and psoas muscle area to height ratio outperformed commonly used frailty indexes in predicting post-esophagectomy mortality and morbidity The index provides a robust picture of patients’ fitness for surgery beyond the qualitative “eyeball” test.

Graphical Abstract

CENTRAL MESSAGE

Quantitative physiologic metrics of physical fitness predict complications after esophagectomy more accurately than standard qualitative frailty indexes.

Esophagectomy Vitality Index (EVI) error vs Fried (FFI) & Modified Frailty (MFI) Indexes.

Graphical Abstract

INTRODUCTION

There are few surgical interventions of greater magnitude than esophagectomy, which is accompanied by major morbidity of 33% and mortality of 3%.¹ Identifying patients at greatest risk for adverse outcomes after esophagectomy is important for caregivers and patients alike. Risk factors for adverse outcomes include “frailty,” a fairly nebulous variable subject to observer bias. As a result, current assessment of vitality and fitness for esophagectomy is crude and imprecise.^2–5

Thus, to characterize patients’ fitness for esophagectomy, objectives of this study were to 1) measure 4 simple physiologic metrics prior to esophagectomy and investigate the degree to which each provides independent information, 2) use these metrics to develop an Esophagectomy Vitality Index and test it as a predictor of 30-day composite outcome, and 3) compare this index’s predictive accuracy against 2 commonly used, but qualitative, frailty indexes: the Fried Frailty Index³ and the Modified Frailty Index.⁴

PATIENTS AND METHODS

Patients

From January 1, 2018, to April 1, 2019, 86 patients were scheduled for esophagectomy at Cleveland Clinic. With Institutional Review Board approval, 77 were enrolled in this prospective study of physiologic metrics for estimating physical fitness for surgery (Figure E1). Patients undergoing esophagectomy for benign diseases (e.g., end-stage achalasia) and those who ultimately did not undergo esophagectomy (e.g., discovery of metastatic disease) were excluded. All patients were evaluated and consented by 1 clinician (AT) in the outpatient clinic during their preoperative visit. Median age was 65 years [25–75th percentiles, 60–72], with 66 (86%) males and a median body mass index of 27 kg/m2 [24–31] (Table 1). Forty-nine patients (65%) had more than a 10-pound weight loss over the previous year, and 65 (84%) and 55 (71%) had undergone neoadjuvant chemotherapy and radiation, respectively.

Table 1.

Patient and Cancer Characteristics and Esophagectomy Details

Variables	na	No. (%) or Median [25th-75th Percentiles]
Demographics
Age (y)	77	65 [60–72]
Male (%)	77	66 (86)
Body mass index (kg/m2)	77	27 [24–31]
Esophageal cancer characteristics
Adenocarcinoma	77	72 (94)
Squamous cell carcinoma	77	5 (6.5)
Clinical T stage	77
T1		8 (11)
T2		19 (25)
T3		50 (65)
T4		0 (0)
Grade	68
Well differentiated		6 (8.8)
Moderately differentiated		36 (53)
Poorly differentiated		26 (38)
Comorbidities
Prior non-esophageal cancer	77	3 (3.9)
Weight loss (>10 lb in past year)	77	49 (64)
Pulmonary function
FEV1 (% of predicted)	76	91 [81–101]
DLCO (% of predicted)	76	81 [70–96]
Cardiac function
Ejection fraction (%)	69	63 [58–68]
Laboratory findings
Albumin (g/dL)	77	4.0 [3.8–4.2]
Creatinine (mg/dL)	77	0.91 [0.76–1.1]
Hemoglobin (g/dL)	77	13 [12–14]
Hematocrit (%)	77	39 [35–42]
Preoperative physical fitness
Weekly METs	77	21 [7.3–46]
CESD Exhaustion Question 1: I felt everything I did was an effort (past week)	77
Rare (<1 day)		47 (61)
Some (1–2 days)		16 (21)
Moderately (3–4 days)		5 (6.5)
Most (5–7 days)		9 (12)
CESD Exhaustion Question 2: I could not get “going” (past week)	77
Rare (<1 day)		61 (79)
Some (1–2 days)		7 (9.1)
Moderately (3–4 days)		5 (6.5)
Most (5–7 days)		4 (5.2)
Preoperative cancer therapy
Chemotherapy	77	65 (84)
Radiotherapy	77	55 (70)
Days from preoperative assessment to surgery		6 [3–11]
Type of esophagectomy for cancer	77
McKeown		42 (55)
Ivor Lewis		14 (18)
Thoracoabdominal		20 (26)

^a.Patients with data available

Key: CESD, Center for Epidemiological Studies Depression scale; CVA, cerebrovascular accident; DLCO, percent predicted diffusing capacity of lung for carbon monoxide; FEV1, forced expiratory volume in 1 second; METs, metabolic equivalents.

Physiologic Metrics

Conduct of study

We measured upper body strength (grip strength), lower body strength and balance (30-second chair sit-stands), muscle mass (psoas muscle area to height ratio), and cardiopulmonary endurance (6-minute walk distance). These metrics were assessed during the preoperative visit immediately preceding the planned operation. Typically, this visit occurred within the week prior to esophagectomy and after neoadjuvant treatment for patients who underwent induction therapy.

Grip strength⁶

Using their dominant hand, patients performed 3 consecutive isometric contractions of a calibrated dynamometer for 5 seconds each time. The measurements were averaged and recorded. For reference, an average dominant-handed grip strength is 43 kg for a 60-year-old man and 25 kg for a 60-year-old woman.⁶

30-second chair sit-stands⁷

Patients were seated on a straight-backed chair with their arms placed across their chest and feet flat on the floor. They were asked to stand up straight to a full standing position while keeping their back straight and arms across their chest. This was performed as many times as the patient could comfortably tolerate over 30 seconds. If a patient was over halfway to a full standing position when 30 seconds elapsed, this was counted as a stand. For reference, a 60-year-old man would be below average if he performed fewer than 14 sit-stands, and a 60-year-old woman would be below average if she performed fewer than 12 sit-stands.⁷

Psoas muscle area⁸

On axial computed tomography imaging, cross-sectional areas of right and left psoas muscles were measured at the inferior edge of the L3 endplate at the level of the iliac crests.⁹ The average of these 2 measurements was recorded and normalized to patient height (cm²/m). This method of normalization was data-driven for this study, chosen over the Psoas Muscle Index ([right psoas area + left psoas area]/height²) because average psoas muscle area to height ratio was found to be a more important predictor.

6-minute walk⁹

Patients were asked to walk at a comfortable pace for 6 minutes, traveling back and forth on a flat 30-meter surface. The distance traveled during this time and any oxygen desaturation events were recorded. If a patient was unable to walk for the full 6 minutes, the time until the patient stopped walking was recorded. For reference, a healthy 65-year-old man can walk 575 m and a healthy 65-year-old woman 550 m in 6 minutes.⁹

Existing Frailty Indexes

Fried Frailty Index³

The Fried Frailty Index categorizes frailty based on 5 qualities: exhaustion, shrinking, low physical activity level, weakness, and slow walking speed (Table E1). Level of exhaustion was determined by asking 2 questions from the Centers for Epidemiologic Studies Depression scale (Table E1).¹⁰ If a patient had 3 or more of these qualities, the patient was labeled “frail,” if 1 or 2, “intermediately frail,” and if none, “not frail.”

Modified Frailty Index⁴

The Modified Frailty Index is an 11-item count of comorbidities from the original 70-item Canada Study of Health and Aging Frailty Index (Table E2).^4,5 To calculate the index in this study, patients were asked about the 11 comorbidities, assigning 1 point to each self-reported comorbidity to generate a score ranging from 0 to 11. The Index was calculated a second time based on the same comorbidities abstracted by full-time registry nurses into the Society of Thoracic Surgeons (STS) General Thoracic Surgery Database. Both the patient-reported and STS-reported indexes were considered in the analysis.

Endpoint

The endpoint was a composite of 30-day mortality and STS-defined complications of delirium, respiratory and gastrointestinal systems (including anastomotic leaks), postoperative length of stay ≥14 days, discharge to facility, and readmission within 30 days of hospital discharge (Table 2). These outcomes were chosen prior to patient recruitment and study analysis.

Table 2.

Postesophagectomy Outcomes Incorporated Into Composite Endpoint (n=77)

Outcomes	No. (%)
30-day mortality	1 (1.3)
Complications a	18 (23)
Pulmonary
Pleural effusion requiring drain	5 (6.5)
Pneumonia	6 (7.8)
Pulmonary embolism	1 (1.3)
Pneumothorax	1 (1.3)
Prolonged intubation	2 (2.6)
Gastrointestinal
Anastomotic leak	3 (3.8)
Requiring treatment	3 (3.8)
Requiring reoperation	2 (2.6)
Conduit necrosis	0 (0)
Esophageal dilatation	0 (0)
Chylothorax	0 (0)
Clostridium difficile infection	2 (2.6)
Delirium	5 (6.5)
Prolonged length of stay (≥14 days)	14 (18)
Discharge to nursing facility	9 (12)
Readmission within 30 days of discharge b	7 (9.2)

^a.Not mutually exclusive.

^b.n=76

Data Analysis

Continuous variables are summarized as median and 25th-75th percentiles and categorical variables as frequency and percentage.

Random forest (RF) classification methodology using the quantile-classifier approach for class-imbalanced data was used to analyze the composite endpoint.¹¹ All computations used open-source random ForestSRCR-software under default settings. Missing data were pre-imputed without outcome information using R-software missForest.¹² Thereafter, 1,000 trees were grown, with each tree constructed using an independent bootstrap sample consisting on average of 63% of the original patients (in-sample bootstrap data). The remaining patients were referred to as out-of-bag (OOB) observations. Each tree and its corresponding OOB observations were used to calculate an OOB (cross-validated) odds and variable importance (VIMP) measure for each of the independent variables (VIMP).¹³ To identify the important candidate variables, a random forest was grown considering the 4 measured metrics: Fried Frailty Index, patient reported and STS-abstracted Modified Frailty Index, and patient and cancer characteristics (candidate variables listed in Appendix E1).

To interpret VIMP and visualize the shape of relations between outcomes and independent variables, partial dependency graphs were produced.¹⁴ The risk-adjusted partial dependency graph provides the forest-averaged prediction for an independent variable across 1,000 trees, adjusting for the predictors in the random forest.¹⁵ Although no formal cutoff analyses were performed, we visually identified inflection points from these partial dependency graphs.

We then created 3 separate imbalanced random forests, each containing the most important and universally applicable predictors from the endpoint random forest described previously with:

1. Four measured metrics

2. Fried Frailty Index

3. Modified Frailty Index, calculated from STS-abstracted comorbidities rather than patient self-reported ones.⁵

The geometric mean error (G-mean), 1 minus the square root of the product of sensitivity and specificity, was calculated for each random forest model. G-mean error is influenced by the number of trees grown in the random forest; however, because the number of trees in each random forest was the same for each of the 3 separate imbalanced random forests, G-mean error estimates are directly comparable. We compared these using the Wilcoxon rank-sum test to identify the most accurate predictive model.

RESULTS

Physiologic Metrics

Median grip strength was 38 kg [31–44], number of sit-stands 11 [10–14], psoas muscle area to height ratio 6.9 cm²/m [6.0–8.2], and 6-minute walk distance 407 m [368–451] (Table 3). Correlation between these metrics was generally weak, with the highest correlation between 30-second sit-stands and 6-minute walk distance (r=0.57) (Figure E2). According to the Fried Frailty Index, the majority of patients (n=56 [73%]) were classified as intermediately frail, 6 (7%) as frail, and 15 as non-frail. The median STS-abstracted Modified Frailty Index was 1 [0–2], and the median patient-reported Modified Frailty Index was 2 [1–3]. On an individual patient basis, most comorbidities other than hypertension and chronic obstructive pulmonary disease (both 83% concordance) were concordant greater than 90% of the time between patient-reported and STS-abstracted data (Table E3). However, only 11 patients (14%) had patient-reported and STS-abstracted comorbidities that were identical.

Table 3.

Physiologic Metrics Measured and Values for Fried Frailty Index and Modified Frailty Index (n=77)

Variable	No. (%) or Median [25th-75th Percentiles]
Physiologic metrics
Grip (kg)
#1	39 [31–44]
#2	38 [30–44]
#3	36 [30–42]
Average	38 [31–44]
30-second chair sit-stands (#)	11 [10–14]
Psoas muscle area (cm2/m)	6.9 [6.0–8.2]
6-minute walk distance (m)	407 [368–451]
Fried Frailty Index components (Table E1)
Shrinking	49 (64)
Exhaustion	19 (25)
Low physical activity	16 (21)
Weakness	12 (16)
Slow walking speed Frailty level	5 (6.5)
None	15 (19)
Intermediate	56 (73)
Frail	6 (7.8)
Modified Frailty Index comorbidities (Table E2)	STS	Patient Reported
Diabetes	22 (29)	21 (27)
Congestive heart failure	0 (0)	4 (5.2)
Hypertension	48 (62)	45 (58)
Transient ischemic attack/cerebrovascular accident	6 (7.8)	5 (6.5)
Dependent functional status	2 (2.6)	2 (2.6)
Myocardial infarction	6 (7.8)	7 (9.1)
Peripheral arterial disease	6 (7.8)	4 (5.2)
Cerebrovascular accident with deficits	2 (2.6)	2 (2.6)
Chronic obstructive pulmonary disease	4 (5.2)	9 (12)
Coronary artery disease	11 (14)	10 (13)
Dementia	2 (2.6)	2 (2.6)
Total score	1 [0–2]	2 [1–3]

Outcomes after Esophagectomy

Twenty-eight patients (36%) developed at least 1 of the composite outcomes (Table 2). Six (7.8%) developed pneumonia, 5 (6.5%) developed delirium, and 3 (3.8%) had anastomotic leaks, with 2 (2.6%) requiring reoperation One patient (1.2%) died within the 30-day perioperative period, 14 (18%) were hospitalized ≥14 days, 9 (12%) were discharged to a nursing facility, and 7 (9.2%) were readmitted within 30 days of hospital discharge.

Predictors of Esophagectomy Outcomes

The 4 measured frailty metrics, in addition to age (Figure E3), smoking pack-years, patient-reported history of transient ischemic attack or cerebrovascular accident (TIA/CVA), percent predicted forced expiratory volume in 1 second (FEV1), neoadjuvant chemotherapy, Fried Frailty Index, and American Society of Anesthesiologists score were the most important predictors of the composite outcome (Figure 1, A). Fried Frailty Index and patient-reported Modified Frailty Index were more important than that calculated from STS-abstracted data (Figure 1, B). Patients with an average grip strength lower than 35 kg and those performing fewer than 10 sit-stands in 30 seconds had a higher predicted risk of developing the composite outcome (Figure 2, A and B), as did patients with a psoas muscle area to height ratio <7 cm²/m (Figure 2, C). Patients who were able to walk a longer distance had a linearly lower predicted risk of developing the composite endpoint without a plateauing effect (Figure 2, D).

Figure 1.

Standardized variable importance (VIMP) for predicting complications. A, Top candidate variables. Note that the prefix Hx represents comorbidities abstracted into our general thoracic surgery database, and the duplicated hypertension variable was patient-reported. B, Esophagectomy Vitality Index variables consisting of grip strength, 30-second chair sit-stands, 6-minute walk distance, and psoas muscle area to height ratio, with Fried Frailty Index (FFI), patient-reported Modified Frailty Index, and STS-Modified Frailty Index. Blue boxes represent important variables with lower 95% confidence interval not extending below 0. Red box represents VIMP value with lower 95% confidence interval extending below 0. Boxes encompass median (line) and 25th and 75th percentile confidence limits, and whiskers 95% confidence limits. Black vertical line at 0.0 VIMP represents the point at which a variable does not contribute predictive ability. Key: ASA, American Society of Anesthesiologists; CESD Q2, Center for Epidemiological Studies Depression scale, question 2; COPD, chronic obstructive pulmonary disease; DLCO, diffusing capacity of lung for carbon monoxide; Esoph, esophageal; FEV1, forced expiratory volume in 1 second; FFI, Fried Frailty Index; MFI, Modified Frailty Index; Neoadj chemo, neoadjuvant chemotherapy; PAD, peripheral arterial disease; Preop, preoperative; STS, Society of Thoracic Surgeons General Thoracic Surgery Database comorbidity definition, TIA/CVA, transient ischemic attack/cerebrovascular accident.

Figure 2.

Risk-adjusted partial dependency plots based on the Esophagectomy Vitality Index estimating the risk of a patient developing the composite outcome after esophagectomy based on A) grip strength, B) 30-second chair sit-stands, C) psoas muscle area to height ratio, and D) 6-minute walk distance. Red circles = risk-adjusted estimates, and blue lines = smooth fit of risk-adjusted estimates.

Esophagectomy Vitality Index versus Fried Frailty Index and Modified Frailty Index

Grip strength, 30-second sit-stands, psoas muscle area, and 6-minute walk distance, in combination with age and self-reported level of exhaustion according to the Centers for Epidemiologic Studies Depression scale questionnaire, comprised the Esophagectomy Vitality Index. Predictions based on the random forest for this index had lower prediction error 32% [31%-33%] than the Fried Frailty Index, 37% [37%-38%; P<.001] (Figure 3, A). Similarly, these predictions had lower error than the STS-abstracted Modified Frailty Index (G-mean error 48% [47%-48%]; P<.001) (Figure 3, B). Compared with the Fried Frailty Index and STS-abstracted Modified Frailty Index, our Esophageal Vitality Index assigned higher predicted risk of the composite outcome for patients who ultimately developed the outcome (Figure 3, red circles). Similarly, it assigned a lower predicted risk of the composite outcome for patients who did not ultimately develop the outcome (Figure 3, blue circles).

Figure 3.

Scattergram of predicted probability of developing the composite outcome and G-mean error based on Esophagectomy Vitality Index (EVI) random forest (RF) versus Fried Frailty Index (FFI) random forest (A) and versus Modified Frailty Index from the STS-based comorbidities (STS MFI) random forest (B). Circles represent each patient with their associated predicted probability for each model and observed outcome. Solid lines are line of identity. Red circles = observed outcome. Blue circles = observed no outcome.

DISCUSSION

Principal Findings

Grip strength, number of 30-second chair sit-stands, psoas muscle area to height ratio, and 6-minute walk distance are simple physiologic metrics, each providing independent information reflecting unique aspects of physical vitality and fitness for esophagectomy. When considered individually, these metrics were among the most important predictors of the poor outcome after esophagectomy for cancer, and when considered together in what we term the Esophagectomy Vitality Index, they appeared to be more informative in predicting an unfavorable outcome following esophagectomy than the Fried Frailty Index or Modified Frailty Index (Figure 4, Graphical Abstract).

Figure 4:

Few surgical interventions are of greater magnitude than esophagectomy, which is accompanied by major morbidity of 33% and mortality of 3%. Risk factors for adverse outcomes include “frailty.” Thus, to characterize patients’ fitness for esophagectomy, this study measured 4 physiologic metrics prior to esophagectomy: grip strength, number of 30-second chair sit-stands, 6-minute walk distance, and psoas muscle area to height ratio. These, plus age, outperformed commonly used frailty indexes in predicting post-esophagectomy mortality and morbidity. The resulting Esophageal Vitality Index provides a robust picture of patients’ fitness for surgery beyond the qualitative “eyeball” test.

Clinical Significance

Risk estimation for any medical intervention depends on the specific outcome being investigated. Mortality, the most important outcome, has become so infrequent for many “high-risk” interventions (including esophagectomy) that risk models are rapidly becoming underpowered and inaccurate. However, as mortality recedes, new adverse outcomes gain importance. For esophagectomy, these include major specific morbidity (e.g., anastomotic leak), length of stay, 30-day readmission, and destination at time of hospital discharge. Although not of equal importance, it has become a pragmatic necessity to formulate composites of these outcomes to provide sufficient power to assess risk, despite loss of specific interpretability.^16,17

Frailty Indexes

Two standard indexes, the Fried Frailty Index³ and the Modified Frailty Index,^2,4,5 represent opposite ends of the spectrum of assessing frailty. The Fried Frailty Index was prospectively developed through the Cardiovascular Health Study,³ broadly categorizing a general population of patients age 65 years or older as frail, intermediately frail, or not frail. Its primary objective was to predict falls, hospitalization, and all-cause mortality within 5 years. It more closely captures what clinicians think of as the “eyeball test.” In contrast, the Modified Frailty Index was developed as a comorbidity index through the American College of Surgeons National Surgical Quality Program.² As a simple tally of patient comorbidities—a measure of disease complexity—it does not measure any physical ability and may be underestimated in large databases due to missing documentation.¹⁸ Neither index was designed specifically for general thoracic surgery patients. Additionally, it appears, based on our findings, that what the patient reports and what is abstracted into quality databases often are not perfectly concordant. This is related in part to incomplete medical record documentation, but, particularly in the case of STS variables, more directly to strict definitions of comorbidities that are necessary to mitigate inappropriate up-coding.

Quantifying Physical Status and Fitness for Esophagectomy

Accurate estimation of patient risk is important in general thoracic surgery, where diseases are often debilitating, and operations such as esophagectomy are of considerable magnitude. Patients undergoing esophagectomy for cancer frequently are malnourished and weakened from prolonged dysphagia and induction therapy regimens, which are associated with decline of physiologic function. A 70-year-old man with well-compensated congestive heart failure and early-stage esophageal cancer who walks 10,000 steps a day may be a lower-risk operative candidate than a 60-year-old previously healthy man with locally advanced cancer who has lost 30 pounds over the past 6 months and subsequently lacks the energy to even walk to the front door. Historically, the eyeball test was used to characterize such patients’ risk for planned intervention. Subjective variables such as posture, ambulation into the office, and attention to the discussion often served as the basis for surgical candidacy. This is almost certainly not reproducible or accurate when screening patients for operative interventions of formidable magnitude.

Our Esophagectomy Vitality Index uses 4 independent objective metrics to assess upper and lower body strength and balance, muscle mass, and cardiopulmonary endurance. These are representative of patients’ real-time physiologic status that do not require costly resources to measure. All study patients performed these metrics within 10 minutes.

Prehabilitation Potential

Frailty is a marker of physiologic fitness and therefore may be modifiable. Although studies have shown that using targeted nutritional and exercise regimens can improve various metrics, their effects on postoperative outcomes are not as well known.^19,20 We have visually identified inflection points in the 4 measured metrics that can serve as tangible target goals for patients to achieve in a prehabilitation program.

Limitations

This is a single-institution prospective study representing only a 15-month experience, but is a contemporary series. Patients studied were a select group who already had demonstrated adequate cardiopulmonary function as assessed by cardiac stress and pulmonary function tests. Our analyses were limited by number of outcome events, and thus we constructed a composite outcome based on frequency and perceived importance of individual postoperative adverse events. Because of this, we could not identify predictors of individual complications. Further, we equally weighted each outcome in the composite for simplicity, without any attempt to hierarchically assign individual weights. If the simple and objective physical status and fitness metrics we measured were added to national quality databases, they may be able to provide more accurate predictions for individual outcomes. It is difficult to know if these metrics are the most predictive frailty ones to use. Others may include nutritional indexes such as recent percentage weight loss, neurocognitive function, or adequacy of social support. Future refined frailty indexes may need to incorporate these as well.

Conclusions

A simple assessment of frailty, which we call the Esophagectomy Vitality Index, consisting of grip strength, number of 30-second chair sit-stands, 6-minute walk distance, and psoas muscle area to height ratio, combined with standard documented clinical variables, outperformed commonly used frailty models in predicting morbidity after esophagectomy for cancer. These objective measures provide a robust picture of a patient’s physical status and fitness for surgery and appear to reflect and look beyond the “eyeball” test.

PERSPECTIVE.

Grip strength, 30-second chair sit-stands, psoas muscle area to height ratio, and 6-minute walk distance each provide a measure of patient fitness. Using them together—the Esophagectomy Vitality Index from this study—predicts a patient’s likelihood of recovery after esophagectomy more accurately than the more commonly used Fried Frailty Index or Modified Frailty Index.

ABBREVIATIONS AND ACRONYMS

FEV1

Forced expiratory volume in 1 second

RF

Random forest for classification