Prognostic impact of myosteatosis on postoperative outcomes and survival in colorectal cancer: a systematic review

Abstract

Purpose

To characterise (1) the definition and measurement methods of myosteatosis, (2) the impact of myosteatosis on post-operative complications (overall, severe complications defined as Clavien-Dindo 3 and above, anastomotic leak, length of stay) in both non-metastatic and metastatic colorectal cancer (CRC) patients and (3) the impact of myosteatosis on long-term survival outcomes in non-metastatic CRC (overall survival, cancer-specific survival, disease-free survival).

Methods

A systematic search of Medline, Embase and Cochrane Central databases was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines in 2024.

Results

A total of 4410 publications were initially retrieved; 54 studies fulfilled the inclusion criteria. A variety of measurement methods of myosteatosis were found. There was good quality evidence to support a negative correlation between myosteatosis and overall survival (OS). Colon cancer was more consistently associated with OS and DFS compared to rectal cancer. Myosteatosis was found to significantly impact severe (Clavien-Dindo 3 and above) postoperative complications in colon cancer. Other postoperative outcomes were limited by inconsistency and an overall paucity in volume and quality.

Conclusion

Myosteatosis has a negative association with OS, with consistent effects observed in colon cancer. There is a negative association with severe postoperative outcomes, particularly in colon cancer. These results are limited by the risk of bias and heterogeneity in measurement methods and statistical analysis. Future work is required to clarify the scan acquisition protocol and diagnostic criteria.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00384-025-05061-y.

Introduction

Colorectal cancer (CRC) is the third most frequently diagnosed cancer worldwide [1]. The prognostication, selection and preparation of patients for surgical and oncological care are important to ensure that we offer patients precision medicine and optimise individual patient survival and quality of life [2, 3]. Surgical management can have severe short-term complications such as infection, haemorrhage, anastomotic leakage and cardio-respiratory complications [4]. Such complications lead to increased morbidity for the patient as well as healthcare costs [5]. Emerging research suggests that patients’ body composition can serve as a prognostic indicator for postoperative recovery and survival.

Myosteatosis, the infiltration of fat within skeletal muscle, is a body composition parameter which increases with aging and has been associated with loss of muscle strength and function. It is thought to precede the development of metabolic disturbances, including insulin resistance and diabetes [6, 7]. Myosteatosis is distinct to sarcopenia (the loss of muscle mass), has independent prognostic effects and has been associated with poorer survival outcomes, prolonged hospitalisation and increased postoperative complications in CRC [8, 9]. Routinely performed computed tomography (CT) scans for oncological staging can provide an opportunistic method for the radiological assessment of myosteatosis in CRC patients.

This systematic review aimed to evaluate the following: (1) The definition and measurement methods of myosteatosis, (2) the impact of myosteatosis on post-operative complications (overall, severe Clavien-Dindo 3 and above, anastomotic leak, length of stay) in both non-metastatic and metastatic CRC patients and (3) the impact of myosteatosis on long-term survival outcomes in non-metastatic CRC (overall survival, cancer-specific survival, disease-free survival) in CRC.

Methods

Search strategy and study selection

This review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. A systematic search was performed on Medline, Embase and Cochrane Central databases from inception up until 31 December 2024. Medical Subject Headings of the National Library of Medicine (MeSH) descriptors related to “colorectal cancer”, “body composition” or “myosteatosis” and “survival” or “postoperative complications” were combined with a broad range of modifiers. The search strategy was developed for MEDLINE and EMBASE and adapted for Cochrane with the aid of an experienced Health Sciences Librarian. The final search strategies are listed in full in the Appendix. Title and abstract screening and full-text review of all texts were conducted independently by two reviewers (CH and JZ), with disagreements resolved through consultation with a third reviewer (JY).

Inclusion criteria for all questions were: (i) patient population of primary colon and/or rectal adenocarcinoma patients, aged 18 and over; (ii) computed tomography images obtained pre-surgery; (iii) measurement of myosteatosis quantified on CT scans; (iv) outcomes of the study must be either postoperative complications and/or long-term survival outcomes, defined as overall survival (OS), recurrence-free survival (RFS) or cancer-specific survival (CSS). When selecting studies analysing survival outcomes, the significant impact of metastasis on prognosis was discussed among all reviewers and agreed to be a confounder [10, 11]. Thus, only studies examining non-metastatic (stage 0–III) CRC were included for this question. For analysis of postoperative outcomes, both non-metastatic and metastatic CRC were included. Studies were excluded if they: (i) omitted data on cancer stage of patients; (ii) lacked pre-operative CT scan acquisition; (iii) CT scans were acquired post-operatively; or (iv) lack of clear description of myosteatosis quantification on CT. Due to the large number of studies retrieved through the search, a best-evidence approach was employed after discussion between reviewers, and only studies with a sample size > 50 were considered for this review.

Data extraction and synthesis

Data extraction was performed by two reviewers (CH and JZ). Extracted study details included authorship, year of publication, country, patient characteristics, treatment regimens including chemo-/radiotherapy regimes, surgical procedures undergone by the patients, the measurement and definition of radiological myosteatosis and outcome data including OS, CSS and RFS. Postoperative outcomes included total complications, complication severity (with severe complications defined as a Clavien-Dindo score of 3 or greater), anastomotic leak and length of stay (LOS). Due to heterogeneity found in the assessment and definition of myosteatosis, patient clinical features including cancer type and treatment regimens and approaches to statistical analyses used among the studies, a meta-analysis was not performed.

The Newcastle–Ottawa Scale was used for assessment of the risk of publication bias of the included studies [12]. CH and JZ independently scored all studies and resolved discrepancies through discussion and escalation to JY where conflicts persisted. Benchmarks for each category were determined through discussion among all reviewers. A clear description of data sources was considered important for all studies. BMI, age, sex, tumour stage and type of surgery were considered important confounding variables, as was smoking, particularly for studies examining postoperative outcomes. A duration of 5 years was considered the standard for sufficient follow-up time in survival outcomes, given that this is the standard for follow-up of CRC internationally.

Results

A total of 4410 publications were retrieved, and after screening and full-text review, a total of 54 studies were found amenable for data extraction (Figure 1).

Fig. 1.

PRISMA flow chart

Study characteristics

Fifty-four studies employing an observational cohort study design were identified, from different regions across Europe, America and Asia (Table 1). Five were prospective, while the remaining were retrospective. Approximately equal gender splits (40–60% male) were found in all studies, and participant age ranged from 40 to 80 s, with a median age of 60–70 being common. The majority of studies specified whether participants had undergone neoadjuvant or adjuvant therapy and the type of surgery (emergency vs elective, open vs laparoscopic vs robot-assisted) performed. Most studies analysed colon and rectal cancer together. Follow-up periods varied from unspecified lengths of time to 10.2 years.

Table 1.

Study characteristics

Study	Country	Study type	Number of participants	Characteristics of subjects (i.e. medium age, gender %)	Chemo- and/or radiotherapy (yes/no)	Type of surgery (i.e. elective vs emergency, open vs laparoscopic)	Measure of myosteatosis (SMD, IMT, IMAC)	Level of measurement (e.g. L3)	Cut-off value for myosteatosis	Muscles measured (i.e. total abdominal muscles vs psoas muscle only)	% myosteatosis
Lee 2023[13]	South Korea	Retrospective cohort study	616	Median (interquartile) age was 64 (55–72) years; male (61.1%)	NO NCRT; some had adjuvant chemotherapy	Elective surgery only (no emergency)	SMD	L3	Martin’s definition	TAMA	32.40%
Lee 2023[14]	USA	Retrospective cohort study	1766	Mean age ± SD: 63.7 ± 14.4 years; 48.8% men	Unknown	Unknown	SMD	L3	Lowest tertile of SMD (ALL: 34.6 HU); and continuous variable	TAMA	58.30%
Kim 2023[15]	South Korea	Retrospective cohort study	474	Age ≥ 65: 58.6%; 54.2% male	Yes, adjuvant chemotherapy only; no, NCRT	Unknown	SMD	L3	Male: < 24.01 HU; female: < 12.50 HU	TAMA	24.30%
Cuijpers 2022[16]	Netherlands	Prospective cohort study	238	Mean age ± SD: 69.3 ± 9.9 years; male 56.3%	Yes, 21.4% had NCRT	Elective surgery only; open and laparoscopic	SMD	L3	Martin’s definition	TAMA	41.20%
Pacquelet 2022[17]	France	Retrospective cohort study	202	Mean age ± SD: 64.8 ± 11.4 [range 29–88]; Male 59.4%	Yes, NCRT	Open and laparoscopic	Intermuscular fat area (IMFA)	4 levels: L2–L3 intervertebral disc; L3; and L3–L4 and L4–L5 intervertebral discs	Continuous variable	TAMA	Not stated
VanVugt 2018 [18]	Netherlands	Retrospective cohort study	816	Median age 79; 53.9% male	Yes, NCRT	Open and laparscopic	SMD	L3	Martin’s definition	TAMA	64.10%
Kemper 2023 [19]	Germany	Retrospective cohort study	207	Mean age ± SD: 65.0 ± 14.5 years; Male 59.4%	Yes, NCRT	laparoscopic or open	SMD	L3	Survival analysis cut-off: median MRA (38 HU); postoperative complications cut-off: continuous variable	TAMA	11.10%
Blackwell 2023[20]	UK	Retrospective cohort study	1122	Median age 70 [range 61–76]; Male 57.7%	Unknown	Elective surgery only	SMD	L3	Martin’s definition	Psoas muscle	Not stated
Schaffler-Schaden 2020[21]	Austria	Retrospective cohort study	85	Median age (IQR) for males and females, respectively: 76 (67.5, 82); 75 (71, 82); male 56.5%,	Yes, chemotherapy	Elective surgery; both laparoscopic and open right colectomy	SMD	L3	Martin’s definition	TAMA	Not stated
Pozzuto 2021 [22]	Brazil	Retrospective cohort study	227	Age > 65: 36.6%, male 44.9%	Yes, adjuvant chemotherapy and NCRT	Emergency and elective surgery	SMD; IMAT	L3	Martin’s definition	TAMA	66.10%
Looijaard 2020[23]	Netherlands	Retrospective cohort study	378	Median age: 73.4 (interquartile range 69.5–78.4); male (60.3%)	Yes, Adjuvant chemotherapy	Elective surgery	SMD; IMAT	L3	No cut-off specified; continuous	TAMA	Not stated
Bimurzayeva 2024[24]	South Korea	Retrospective cohort study	499	No age statistics given for whole cohort; male 57.7%	No NCRT	Unknown	SMD	L3	MALE: 34.8 HU; FEMALE: 33.9 HU	TAMA	Not stated
Xiao 2023[25]	China	Retrospective cohort study	191 Rectal cancer	Mean age ± SD: 73.60 ± 8.81 years; 66.0% male	Yes, CRT	Open and laparscopic	SMD	L3	Continuous variable	Psoas muscle only	Not stated
Tankel 2020[26]	Israel	Retrospective cohort study	185	Age > 75: 31.9%; 49.2% male, 50.8% female	Yes, adjuvant or NCRT	Laparoscopic right or extended right hemicolectomy	HUAC (measured as: ([right average HU*right psoas area]/[TPA] + [left average HU*left psoas area]/[TPA])/2)	Upper border of L3	Lowest gender specific quartile (male, 44.5; female, 44.9)	Psoas muscle	23.80%
Chung 2020[27]	South Korea	Retrospective cohort study	167	Age ≥ 65: 31.7%; 59.9% male, 40.1% female	Yes, adjuvant chemotherapy. No NCRT	Elective	SMD	L3	Martin’s definition	TAMA	Not stated
Giani 2020[28]	Italy	Retrospective cohort study	173	Median age 71 (61–77); Male 64.2%	Yes- NCRT	Open and laparoscopic; anterior resection, Miles procedure or Hartmann’s procedure	Intramuscular fat area ratio (IMFAR = IMFA/TMA)	L3	Lowest gender specific quartile (male, 0.28 cm2/m2; female, 0.37 cm2/m2)	TAMA	25.00%
Giani 2022[29]	Italy	Retrospective cohort study	129	Median age 72 years (IQR 62–78); male 68.2%	Yes, adjuvant chemotherapy only	Open or laparoscopic	IMFAR = intramuscular fat area/total muscle area	Abdomen (unspecified level)	Last quartile of dataset (unspecified value)	TAMA	64.10%
Carcamo 2021 [30]	Chile	Retrospective cohort study	359	Mean age ± SD: 64 ± 14.12 years; 193 men (54%) and 166 women (46%)	Yes, adjuvant chemotherapy	Open and laparoscopic	SMD	L3-L4	Martin’s definition	TAMA	22.30%
Ferreira 2024[31]	Brazil	Retrospective cohort study	257	Mean age ± SD 62.5 ± 12.1 years; 49.8% male	Yes	N/A	SMD = mean HU of both legs	Level of the lower margin of the lesser trochanter of the femur	Median value of sample’s data distribution	Quadriceps, pectineus, tensor fasciae latae, iliopsoas, sartorius, adductor longus, adductor brevis, adductor minimus, adductor magnus, gracilis, semitendinosus and gluteus maximus	33%
Kotek 2024[32]	Czech Republic	Retrospective cohort study	206	Median (IQR) age 67 (59–73); 65.5% male	Yes, NCRT	Elective only; robotic, laparoscopic, open	SMD	L3	Continuous variable	Psoas muscle	N/A
Pekarova 2021 [33]	Czech Republic	Retrospective cohort study	112	Mean age ± SD: 68.3 ± 10.4; male 66.0%, female 34%	Yes, 36.2% of rectal cancer cases underwent neoadjuvant chemoradiotherapy	Laparascopic surgery or open	SMD	L3	No cut-off specified; continuous variable	Psoas muscle only	Not stated
Benedek 2021[34]	Romania	Prospective cohort study	51	Median age 65 years (interquartile range (IQR), 56–71); male, 62%	Unknown	Unknown	SMD	L3	Continuous variable	Psoas muscle	Not stated
Herrod 2019[35]	UK	Retrospective cohort study	169	Mean age ± SD: 68 ± 11 years; Male 53.8%	Yes	Unknown	SMD	L3	Whole cohort, SMD ≤ 44.5 HU; anastomosis group, SMD ≤ 43.5 HU	Psoas muscle only	30.20%
Gao 2025 [36]	China	Retrospective cohort study	317	Median age (IQR) 66 (58–72) years; 64.7% male	Not specified	Elective only; laparoscopic or open	IMAC = mean CT value of ROI of multifidus muscle (HU)/mean CT value of ROI of subcutaneous fat (HU)	L3	Sex-specific lowest 25th percentile (34.76 HU for males and 31.82 HU for females)	Psoas muscle	24.9
Parnasa 2025[37]	Israel	Retrospective cohort study	843	55.0% male	Unknown	Laparoscopic and open, elective and emergency included	SMD	L3	Continuous variable	Psoas muscle	N/A
Okugawa 2018[38]	Japan	Retrospective cohort study	308	Age > 67: 48.4%; male 59.4%	Yes, adjuvant chemotherapy	Open and laparoscopic	IMAC = mean CT value of ROI of multifidus muscle (HU)/mean CT value of ROI of subcutaneous fat (HU)	Superior aspect of L4	Sex-specific median value (male: IMAC: − 0.36; female: IMAC: − 0.24)	Multifidus muscle	49.70%
Kusunoki 2021[39]	Japan	Retrospective cohort study	471	Median age: 68 years (range, 27–94 years); 59.7% male	Yes, Adjuvant chemotherapy	Unknown	IMAC = mean CT value of the ROI of the multifidus muscle (HU)/mean CT value of the ROI of the subcutaneous fat (HU); mIMAC = mean CT value of the ROI of the multifidus muscle (HU) − mean CT value of the ROI of the subcutaneous fat (HU)	L4	Male: IMAC: − 0.368 and mIMAC: 143.13; female: IMAC: − 0.433 and mIMAC: 133.06	Multifidus muscle	IMAC, 61.8%; mIMAC, 48.0%
Margadant 2016[40]	Netherlands	Retrospective cohort study	373	Median age: 78 years (IQR 75–82 years); male 48.5%	Yes, adjuvant and NCRT	Laparoscopic or open	HUAC (measured as: ([right average HU*right psoas area]/[TPA] + [left average HU*left psoas area]/[TPA])/2)	L3	Male: ≤ 22.0 HU/mm2; female: ≤ 23.5 HU/mm2; as well as continuous variable	Psoas muscle only	24.70%
Souwer 2020[41]	Netherlands	Retrospective cohort study	174	Mean age ± SD: 78.0 years ± 5.1; 51.1% Male	Yes, Adjuvant chemotherapy and NCRT	All elective; laparoscopic 65.5%	SMD	L3	Martin’s definition	TAMA	87.40%
Chen 2022[42]	China	Retrospective cohort study	921	Median age: 73 years (IQR 10); male 60.8%	Unknown	Laparoscopic or open; no emergency surgeries	SMD	L3	Male, 38.5 HU; female, 28.6 HU	TAMA	68.70%
Boer 2016[43]	USA	Retrospective cohort study	1630	Mean age ± SD: 64.0 ± 11.3 years; male 44.4%	Unknown	Unknown	SMD	L3, L4 superior, L4 inferior	Xiao’s definition	TAMA	Not stated
Berkel 2019[44]	Netherlands	Retrospective cohort study	99 LARC	Median age: 66 (range 40–81); 53.5% males	Yes, NCRT	Open and laparoscopic	SMD	L3	Continuous variable	TAMA	Not stated
Aro 2020[45]	Finland	Retrospective cohort study	348	Mean age ± SD: 68 ± 11.5 years; 182 men (52%) and 166 women (48%)	Yes, adjuvant and/or neoadjuvant CRT	Emergency and elective; open and laparoscopic	SMD	L3	Martin’s definition	TAMA	31.20%
McGovern 2023[46]	Scotland	Retrospective cohort study	1002	Age ≥ 65: 65.6%; male 55.3%	Yes, NCRT	Unknown	SMD	L3	Martin’s definition	TAMA	Not stated
Van der Krof 2018 [47]	Netherlands	Prospective cohort study	63	Mean age ± SD: 69 ± 10.5 years; male 61.9%	Yes, neoadjuvant therapy	Elective surgery only; laparoscopic and open	SMD	L3	Median value as cut-off (low: < 34.1 HU; high: ≥ 34.1 HU)	TAMA	50%
Almasaudi 2019[48]	Scotland	Retrospective cohort study	741	Age ≥ 65: 64.2%; 55.3% male	Yes, NCRT	Laparoscopic or open	SMD	L3	Martin’s definition	TAMA	62.50%
Xiao 2020[49]	US	Retrospective cohort study	1630	Mean age ± SD: 64.0 ± 11.3 years, 44.4% were male	Unknown	Laparoscopic and endoscopic polpectomy only	SMD	L3	Xiao’s definition	TAMA	30.40%
Malietzis 2016[50]	UK	Prospective cohort study	805	Median age of 69 (61–77) years; Male 58.6%	Yes, NCRT and CRT	Laparoscopic or open	SMD	L3	Martin’s definition	TAMA	77.60%
Martin 2018[51]	UK and Canada	Retrospective cohort study	1139	Mean age ± SD: 66.6 ± 11.9 years; male 60.5%	Unknown	Laparoscopic or open; no emergency	SMD	L3	Male: < 50 y: < 42.0; 50–59 y: < 37.6; 60–69: < 32.9; 70–79: < 29.7; ≥ 80: < 28.1; female: < 50 y: < 39.6; 50–59 y: < 36.0; 60–69: < 31.4; 70–79: < 27.8; ≥ 80: < 24.8	TAMA	50%
vanRees 2021 [52]	Netherlands	Retrospective cohort study	227	Age ≥ 70: 32.2%; male 40.5%	Yes, neoadjuvant chemo- and/or radiotherapy therapy for most patients	Total (111) or posterior (116) pelvic exenteration surgery	SMD	L3	Martin’s definition	TAMA	57.20%
Hopkins 2019[53]	Canada	Retrospective cohort study	968	Mean age ± SD: 65.8 ± 11.8 years; 589 men (61%) and 379 women (39%)	Yes, Adjuvant CRT	Unknown	SMD	L3	Martin’s definition	TAMA	61.10%
Sueda 2018[54]	Japan	Retrospective cohort study	211	Age ≥ 65: 50.7%; 134 men (63.5%)		Elective surgery only; laparoscopic and open	SMD	L3	Martin’s definition	TAMA	52.10%
McSorley 2018[55]	Scotland	Retrospective cohort study	322	Age ≥ 65: 67%; 174 men (54%) and 148 women (46%)	Unknown	Elective surgery	SMD	L3	Martin’s definition	TAMA	57.80%
vanBaar 2018 [56]	Netherlands	Retrospective cohort study	1681	Mean age ± SD: 67.7 ± 10.3 years; Men 59.4%	Unknown	Unknown	SMD	L3	Men: BMI < 25 kg/m2: 36.4 HU, BMI ≥ 25 kg/m2: 31.6 HU; Women: BMI < 25 kg/m2: 31.1 HU, BMI ≥ 25 kg/m2: 29.3 HU	TAMA	38.50%
Dolan 2019[57]	Scotland	Retrospective cohort study	650	Age > 65: 64.0%; male 54.5%	Unknown	Elective surgery only	SMD	L3	Xiao’s definition	TAMA	52.50%
Aro 2022[58]	Finland	Retrospective cohort study	222	Age ≥ 70: 57.2%; male 52.7%	No NCRT	Unknown	SMD	L3	Martin’s definition	TAMA	30.20%
Koh 2022[59]	South Korea	Retrospective cohort study	1015	Age ≥ 70: 30.8%; Male 59.0%	No NCRT	Unknown	SMD	L3	Martin’s definition	TAMA	Not stated
Cho 2024[60]	South Korea	Retrospective cohort study	1053	Mean age ± SD: 62.8 ± 11.8 years; male [59.1%]	Yes, adjuvant chemotherapy	Unknown	SMD	L3	Martin’s definition	TAMA	29.40%
Kim 2023[61]	South Korea	Retrospective cohort study	906	Median age: 64 (IQR 55–72) years: 59.7% male	Yes, adjuvant chemotherapy	Unknown	SMD	L3	Martin’s definition	TAMA	Not stated
Golder 2022[62]	Scotland	Retrospective cohort study	1146	Age > 65: 36.8%; Male 51.4%	Yes, adjuvant chemotherapy	Either elective or emergency	SMD	L3	Xiao’s definition	TAMA	62.00%
Zhou 2024[63]	China	Retrospective cohort study	327	Mean ± SD age 60.9 ± 10.6 years, 62.7% female	NCRT excluded	Open and laparoscopic	SMD	Unspecified	Median value of sample’s data distribution (unspecified value)	Unspecified	49%
Renman 2023 [64]	Sweden	Retrospective cohort study	519	Median age: 66.5 years (range 41.5–85.6); male 50.7%	Unknown	Unknown	SMD	L3	Lowest sex-specific tertile of SMD (male, 38.5 HU; female, 36.1 HU)	TAMA	31.40%
Yokoi 2022[65]	Japan	Retrospective cohort study	341	Age ≥ 70: 44.9%; 58.7% men	Yes, NCRT and CRT	Unknown	SMD	L3	Male: < 35.0 HU; female: < 21.7 HU	TAMA	Not stated
Wei 2024[66]	Australia	Retrospective cohort study	214	Mean age ± SD: 62.4 ± 12.7 years; male 70.6%	Yes, NCRT	Open and laparoscopic	IMAT volume, mean IMAT density, SMD	L1–L5	Continuous variable	TAMA (for SMD)	N/A

IMFAR, intramuscular fat area ratio, measured as IMFA/TMA, intramuscular fat area/total muscle area; HUAC, measured as ([right average HU*right psoas area]/[total psoas area] + [left average HU*left psoas area]/[total psoas area])/2); IMAC, intramuscular adipose content, measured as mean CT value of the ROI of the multifidus muscle (HU)/mean CT value of the ROI of the subcutaneous fat (HU); mIMAC, modified intramuscular adipose content, measured as mean CT value of the ROI of the multifidus muscle (HU) − mean CT value of the ROI of the subcutaneous fat (HU); IMFA, intermuscular fat area, measured as fat contained in the spaces between the muscles of the abdominal wall that does not belong to the VFA or SFA compartments. Martin’s definition: body mass index (BMI) < 25 kg/m²: HU < 41; BMI ≥ 25 kg/m²: HU < 33. Xiao’s definition: males, < 35.5 HU; females, < 32.5 HU. SD standard deviation, IQR interquartile range, NCRT neoadjuvant chemotherapy, SMD skeletal muscle density, HU Hounsfield units, TAMA total abdominal muscle area

Measurement and definition of myosteatosis

Myosteatosis was evaluated on preoperative CT scans up to 3 months before surgery. The use of intravenous contrast for these CT scans was specified in 20 of the 54 studies included [13–32].

The most common method for detecting myosteatosis on CT scans was the use of skeletal muscle radiodensity (SMD), which measures the mean radiodensity of tissue areas of −29 to 150 HU in attenuation (Table 1). The most common area examined was the total abdominal muscular area at the level of L3 or L4 (37 of 54 studies). The majority of studies determined the cut-off for myosteatosis by Martin’s definition (< 33HU in men and women of BMI ≥ 25 and < 41 for men and women with BMI < 25) (20 out of 37) and Xiao’s definition (men, < 35.5HU; women, < 32.5HU) (four out of 37) [67, 68]. Figure 2 shows the schematic representation of types of fatty infiltration into muscle (Fig. 2a) and schematic representation of CT-based measurements of myosteatosis at the level of L3 (Fig. 2b).

Fig. 2.

Evaluation of radiologically defined myosteatosis. a Schematic representation of types of fatty infiltration into muscle including intermuscular or extrafascial adipose (D) which is external to the muscle and its fascia (A); perimuscular or subfascial (E) which exists within a muscle external to individual muscle fascicles (B); intramuscular adipose (G) which exists within fascicles adjacent to myocytes (C); and intramyocellular lipid droplets (F) which exist within individual myocytes (C). b Schematic representation of CT-based measurements of myosteatosis at the level of L3 including in panel A, where total abdominal musculature is highlighted in blue and represents the total area from which the average radiological attenuation is measured to generate skeletal muscle density (SMD) measurements. Areas of intermuscular adipose infiltration are highlighted in yellow; the area of these regions may be utilised for calculations of intermuscular fat infiltration or IMAT

Ten studies measured myosteatosis based on examination of the psoas muscles. Seven out of the ten studies evaluated SMD of psoas muscles [20, 25, 32–37], while the other three studies utilised HUAC as the index measurement of myosteatosis. For the studies utilising SMD, the majority of studies treating the variable as a continuous variable in their analysis without any specification of a cut-off value. HUAC was measured as:

$$\frac{(RightpsoasaverageHU\times rightpsoasarea)/(LeftpsoasaverageHU\times leftpsoasarea)}{Rightpsoasarea+leftpsoasarea}$$

All three of the studies utilising HUAC had different sex-specific cut-off values for myosteatosis.

One study by Ferreira et al. utilised the mean SMD of both thighs at the level of the lesser trochanter of the femur as the measure for myosteatosis.

Other studies utilised segmented inter-/intramuscular adipose tissue area (IMAT), with tissue areas of radiological attenuation between −190 and −30 HU defined as IMAT. However, among these studies, the cut-off volume for myosteatosis was heterogenous with most studies opting to utilise analyses of variance.

Variants of IMAT such as inter-/intramuscular fat area ratio (IMFAR) and inter-/intramuscular adipose tissue content were also used as index measurements for myosteatosis. Intramuscular adipose tissue content (IMAC) was described by Okugawa and Kusunoki as a measurement involving the attenuation value (in Hounsfield Units) of the multifidus muscle, divided by the mean attenuation of adipose tissue, adapted from Kitajima’s measurements of fat deposition [38, 39, 69]. Giani et al.’s two studies in Italy measured myosteatosis using IMFAR, defined as intramuscular fat area/total muscle area [28, 29].

Wei et al.’s 2024 study was the only study out of the 54 to utilise SMD measurements over multiple levels (L1-L5) and measured IMAT volume and density rather than just cross-sectional area across L1-L5 [66]. However, cut-off values for IMAT volume and density to diagnose myosteatosis were not proposed. Furthermore, Lee’s 2023 and Wei’s 2024 studies were the only two studies to use fully automated artificial intelligence body composition algorithms to segment and quantify different body composition parameters, including IMAT area and volume and SM area and density [13, 68].

Postoperative complications

Overall complications

Ten studies investigated the association between myosteatosis and overall postoperative complications in colorectal cancer, with only four of these studies finding a significant correlation [16, 18, 24, 36, 41, 42, 45–48]. Two out of the four studies (Van Vugt 2018 and Gao 2025) were deemed to be good quality based on the NOS [18, 36]. Van Vugt’s prospective cohort study utilised SMD TAMA as the measurement for myosteatosis, finding 51.7% of the low SMD group developed postoperative complications compared to 38.5% for those with normal SMD (p < 0.001) [18]. Gao utilised HUAC measurement based on psoas muscles and demonstrated a hazard ratio (HR) of 1.8 (95% CI 1.0–3.3, p = 0.039) [36]. In contrast, two other studies that were considered good quality based on the NOS found no significant association [24, 45].

Four studies investigated the association between myosteatosis and overall postoperative complications in colon cancer only, with half of the studies finding a significant correlation [21, 26, 43, 49]. Boer et al. utilised SMD of TAMA (Xiao’s definition) and Tankel utilised HUAC [26, 43]. However, both studies were deemed to be poor quality based on the NOS; the other two studies on colon cancer, deemed good quality, found no significant association when using SMD of TAMA (Martin and Xiao’s definitions) [21, 49].

Overall, there was a paucity of studies investigating myosteatosis and overall postoperative complications in rectal cancer patients. Out of the three studies investigating the association between myosteatosis and overall postoperative complications in rectal cancer, only one study found a significant association, utilising SMD of TAMA (no cut off; continuous variable); however, this achieved a poor quality rating on NOS [17, 44, 66].

Severity of complications

Fourteen studies investigated the association between myosteatosis and severe postoperative complications (defined as Clavien-Dindo $\ge$ 3) in colorectal cancer [16, 18, 20, 23, 34–37, 40–42, 45, 50, 51]. Five of these studies were assessed to be of good quality, including three which found a significant correlation [18, 23, 35, 36, 40]. Two of the studies (Van Vugt 2018 and Looijaard 2020) utilised SMD based on TAMA, while Gao’ s study utilised psoas HUAC [18, 23, 36]. HRs of these three studies ranged from 1.87 (95% CI 1.01–3.46, p = 0.045) to 3.8 (95% CI 1.5–9.6, p = 0.005).

Out of the three studies investigating the association between myosteatosis and severe postoperative complications in colon cancer only, all three found a significant association [26, 33, 49]. Xiao used SMD of TAMA, while Tankel utilised HUAC of psoas muscles, while Pekarova used SMD of psoas muscles. However, only one study (Xiao 2020) was deemed to be good quality on NOS. Xiao’s study found a HR of 2.41 (95% CI 1.44–4.04, p < 0.05) [49].

Out of the five studies investigating the association between myosteatosis and severe postoperative complications in rectal cancer only, three found a significant association [25, 32, 33, 44, 52]. However, all three were deemed to be poor quality according to NOS.

Anastomotic leak (AL)

Eight studies investigated the association between myosteatosis and anastomotic leak in colorectal cancer, with only one study found to be of good quality [19, 20, 35, 40, 41, 45, 48, 50]; Aro et al. utilised the SMD of TAMA (Martin’s definition) and found that there was no significant association between myosteatosis and the risk of anastomotic leak [44]. Of the eight studies that found a significant correlation, two studies measured the SMD of psoas muscles, one utilised the SMD of TAMA and one used the HUAC of psoas muscles [19, 20, 35, 40]. None was deemed good quality according to NOS.

Overall, there was a paucity of studies investigating myosteatosis and AL in colon or rectal cancer only. Out of the two studies investigating the association between myosteatosis and anastomotic leak in colon cancer only, both found a significant association, while none of the three studies investigating the association between myosteatosis and anastomotic leak in rectal cancer found a significant association [28, 32, 33, 49].

Length of stay (LOS)

Ten studies investigated the association between myosteatosis and length of stay (LoS) in colorectal cancer, with four good quality studies identified [18–20, 23, 37, 40, 42, 45, 50, 51]. Of these, Van Vugt et al. found a correlation between myosteatosis, measured by SMD of TAMA (Martin’s definition), and longer LoS (median LoS 8 days, IQR 6–14, compared with 7 days IQR 5–10, p < 0.001). Other good quality studies did not demonstrate a significant correlation between myosteatosis and LoS [23, 37, 45].

Overall, there was a paucity of studies investigating myosteatosis and LoS in colon or rectal cancer only.

Among two studies focussing on colon cancer, Xiao et al.’s research was deemed to be of good quality and found an increased risk of prolonged LoS in patients with myosteatosis (HR 1.39, 95% CI 1.05–1.84) [49]. Only one study investigated rectal cancer populations only, finding higher SM density of TAMA was an important independent prognostic factor for better LoS in rectal cancer patients (HR 1.58, 95% CI 1.03–2.43, p = 0.036) [66]. Results from the studies investigating the association between myosteatosis and postoperative complications are presented in Table 2.

Table 2.

Association between myosteatosis and post-operative complications in non-metastatic and metastatic CRC

Author	Overall	Severity (i.e. Clavien-Dindo (CD) =/> 3)	Anastomotic leak	Length of stay (LOS)
Colorectal cancer
VanVugt (2018) [18]	Myosteatosis vs non-myosteatosis: 51.7% vs 38.5%, p < 0.001	HR 1.87, 95% CI 1.01–3.46, p = 0.045	-	Myosteatosis vs non-myosteatosis: median [IQR], days 8 [6 - 14] vs 7 [5 - 10], p < 0.001
Kusunoki (2021) [39]	-	-	-	-
Aro (2020) [45]	No significant association	No significant association	No significant association	No significant association
Souwer (2020) [41]	Myosteatosis vs non-myosteatosis: 49% vs 28%, p = 0.046	No significant association	No significant association	-
McGovern (2023) [46]	No significant association	-	-	-
Bimurzayeva (2024) [24]	No significant association	-	-	-
Looijaard (2020) [23]	-	Non-myosteatosis compared to myosteatosis: HR 0.684, 95% CI 0.486–0.962, p < 0.05	-	No significant association
Blackwell (2023) [20]	-	No significant association	HR 3.13, 95% CI 1.05–9.33; p = 0.04	No significant association
Van der Kroft (2018) [47]	No significant association	-	-	-
Benedek (2021) [34]	-	No significant association	-	-
Kemper (2023) [19]	-	-	Every 10HU increase associated with decreased odds of 47%, 95% CI 7.7%–103%, p = 0.017	HR for the event hospital discharge for non-myosteatosis compared to myosteatosis group: HR 1.25, 95% CI: 1.05–1.48, p = 0.011
Chen (2022) [42]	HR 1.478, 95% CI 1.040–2.099, p = 0.029	No significant association	-	Myosteatosis vs non-myosteatosis: median [IQR], days 12 [7] vs 11 [6], p < 0.001
Almasaudi (2019) [48]	Male, no significant association; female, no significant association	-	Male: No significant association; Female: No significant association	-
Margadant (2016) [40]	-	HR 1.84, 95% CI 1.11–3.06, p = 0.019	Myosteatosis vs non-myosteatosis: 23.3% vs 11.7%, p = 0.016	Myosteatosis vs non-myosteatosis: median (min–max), days 13 (2–126) vs 10 (1–130), p = 0.025
Malietzis (2016) [50]	-	No significant association	No significant association	Myosteatosis vs non-myosteatosis: median [IQR], days 7 [5 - 11] vs 6 [4 - 9] days respectively, p = 0.034
Cuijpers (2022) [16]	No significant association	No significant association	-	-
Martin (2018) [51]	-	No significant association	-	No significant association
Herrod (2019) [35]	-	HR 6.33, 95% CI 1.6–24.24, p = 0.007	HR 14.37, 95% CI 1.3–150.0, p = 0.026	-
Gao (2025) [36]	HR 1.8, 95% CI 1.0–3.3, p = 0.039	HR 3.8, 95%CI 1.5–9.6, p = 0.005	-	-
Parnasa (2024) [37]	-	No significant association	-	No significant association
Colon cancer
Schaffler-Schaden (2020) [21]	No significant association	-	-	-
Boer (2016) [43]	L3 level: HR 0.912, 95% CI 0.863–0.964, p = 0.001 L4 (superior) level: HR 0.918, 95% CI 0.871–0.967, p = 0.001; L4 (inferior) level: 0.922, 95% CI 0.875–0.972, p = 0.003	-	-	-
Xiao (2020) [49]	No significant association	HR 2.41, 95% CI 1.44–4.04, p < 0.05	HR 1.67, 95% CI 1.02–2.74, p < 0.05	LOS ≥ 7 days: HR 1.39, 95% CI 1.05–1.84, p < 0.05
Tankel (2020) [26]	HR 1.68, 95% CI 0.60–1.16, p = 0.009	Myosteatosis significantly associated with CD =/> 3 (p = 0.015)	-	No significant association
Pekarova (2021) [33]	-	PD was significantly lower in the Cl-Di III-V cohort compared to the Cl-Di 0-II cohort (Mean ± SD: 42.67 ± 6.52 vs. 40.11 ± 7.57 HU, p = 0.002)	PD was significantly lower in patients with anastomotic leakage compared to those without leak (Mean ± SD: 31.40 ± 7.50 vs. 42.73 ± 6.37, p = 0.034)	-
Rectal cancer
Giani (2020) [28]	-	-	No significant association	-
Berkel (2019) [44]	Non-myosteatosis compared to myosteatosis: HR 0.91; 95% CI 0.85–0.97, p = 0.003	Non-myosteatosis compared to myosteatosis: HR 0.89, 95% CI 0.83–0.96; p = 0.002	-	-
Xiao (2023) [25]	-	Non-myosteatosis compared to myosteatosis: HR 0.94, 95% CI 0.88–0.99, p = 0.04	-	-
vanRees (2021)	-	No significant association	-	-
Pacquelet (2022) [17]	No significant association	-	-	-
Pekarova (2021) [33]	-	PD was significantly lower in the Cl-Di III-V cohort compared to the Cl-Di 0-II cohort (Mean ± SD: 44.08 ± 5.86 vs 43.03 ± 5.70 HU, p = 0.016)	No significant association	-
Kotek (2024) [32]	-	No significant association	No significant association	-
Wei (2024) [66]	No significant association	-	-	SMD: HR 1.58, 95% CI 1.03–2.43, p = 0.036

HR hazard ratio, CI confidence interval, SD standard deviation, IQR interquartile range, HU Hounsfield unit

Survival

Overall survival

Twenty-three studies investigated the association between myosteatosis and overall survival in colorectal cancer [14, 18, 19, 23, 24, 30, 31, 38, 41, 42, 45, 50, 53–61, 63, 64]. Twenty out of the twenty-three studies utilised SMD of TAMA to measure myosteatosis. Fifteen of these studies found a significant correlation, and fourteen of these studies measured SMD of TAMA, with the majority using Martin’s criteria as the cut-off value [16, 28, 39, 43, 48, 51–53, 56–58]. One study (Ferreira, 2024) measured myosteatosis using SMD of the bilateral thighs. However, no cut-off value was specified, as the value was treated as a continuous variable. Cho et al.’s study found a significant association between myosteatosis and OS in the group 51–74 years (non-myosteatosis compared to myosteatosis group HR 0.63 (0.41−0.96), p = 0.033). However, no significant association was found in the age group ≥ 75 years. Only five out of these fifteen studies were deemed to be of good quality according to NOS. The hazard ratios ranged from 1.54 (95% CI 1.19–1.98, p = 0.001) in Hopkins et al.’s study, which utilised SMD of TAMA (Martin’s definition), to 2.43 (95% CI 1.327–4.449, p = 0.004) in Birmurzayeva’s study, which utilised SMD of TAMA and sex-specific cut-offs [24, 53].

Four out of the five studies investigating the association between myosteatosis and overall survival in colon cancer only found a significant correlation (HRs ranging from 1.44, 95% CI 1.12–1.85, p < 0.05; to 1.90, 95% CI 1.84–3.04, p = 0.008) [15, 22, 27, 49, 62]. All four studies utilised SMD of TAMA to quantify myosteatosis, with different cut-off values utilised by each study. Three out of these four studies were deemed to be good quality according to NOS.

One out of the three studies investigating the association between myosteatosis and overall survival in rectal cancer only found significant associations. Interestingly, Wei’s 2024 study found that although higher SMD was an important independent prognostic factor for better OS (HR 0.24, 95% CI 0.06–0.86, p = 0.029), higher IMAT volume and density, which represent increased myosteatosis, were also associated with improved OS (IMAT volume: HR 0.13, 95% CI 0.03–0.59, p = 0.008; IMAT density: HR 0.26, 95% CI 0.10–0.68, p = 0.006) [66].

Cancer-specific survival

Nine studies investigated the association between myosteatosis and cancer-specific survival (CSS) in colorectal cancer, with only three of these studies finding a significant correlation [18, 30, 38, 53–56, 58, 64]. All three of the studies quantified SMD of TAMA and defined myosteatosis using Martin’s criteria. Only one of the three studies was deemed good quality (HR, 1.42; 95% CI, 1.04–1.93, p = 0.026) [53].

No studies investigated CSS in colon or rectal cancer only.

Disease-free survival

Fourteen studies investigated the association between myosteatosis and disease-free survival in colorectal cancer [13, 18, 24, 30, 38, 45, 50, 53, 54, 56, 58, 61, 63, 65]. Six studies in total found a significant correlation between myosteatosis and DFS; all utilised SMD of TAMA to measure myosteatosis, but the cut-off value for myosteatosis varied among the studies. Only two out of these six studies were deemed to be of good quality according to NOS; the other four did not find a significant correlation.

Both studies investigating the association between myosteatosis and disease-free survival in colon cancer found a significant correlation, using SMD of TAMA and different sex-specific cut-offs [15, 22]. One out of the three studies investigating the association between myosteatosis and overall survival in rectal cancer found significant associations. Wei’s 2024 study found that higher SMD was an important independent prognostic factor for better DFS (HR 0.45, 95% CI 0.2–0.99, p = 0.048). However, no significant association was found between IMAT volume and density and DFS. However, these studies were limited by their NOS rating. Results from the studies investigating the association between myosteatosis and survival outcomes in non-metastatic CRC are presented in Table 3.

Table 3.

Association between myosteatosis and survival outcomes in non-metastatic CRC

Author	Overall survival	Cancer-specific survival	Disease-free/recurrence-free survival
Colorectal cancer
Carcamo (2021) [30]	No significant association	No significant association	No significant association
VanVugt (2018) [18]	No significant association	No significant association	No significant association
Hopkins (2019) [53]	HR 1.54, 95% CI 1.19–1.98, p = 0.001	HR, 1.42; 95% CI 1.04–1.93, p = 0.026	No significant association
Aro (2020) [45]	HR 1.6, 95% CI 1.04 to 2.56, p = 0.034	-	No significant association
Sueda (2018) [54]	HR 2.94, 95% CI 1.32–7.17, p < 0.01	HR 3.37, 95% CI 1.38–9.43, p < 0.01	HR 2.73, 95% CI 1.41–5.53, p < 0.01
McSorley (2018) [55]	HR 2.29, 95% CI 1.38–3.81, p = 0.001	HR 2.11, 95% CI 1.14–3.92, p = 0.017	-
Souwer (2020) [41]	No significant association	-	-
vanBaar (2018)	HR 1.91, 95% CI 1.53–2.38 p < 0.05	No significant association	HR 1.68, 95% CI 1.14–2.47, p < 0.05
Renman (2023) [64]	No significant association	No significant association	-
Bimurzayeva (2024) [24]	HR 2.430, 95% CI 1.327–4.449, p = 0.004	-	HR 2.049, 95% CI 1.129–3.718, p = 0.018
Looijaard (2020) [23]	No significant association	-	-
Dolan (2019) [57]	HR 1.42, 95% CI 0.98–2.05, p = 0.061	-	-
Aro (2022) [45]	HR 1.81, 95% CI 1.00–3.26, p = 0.048	No significant association	No significant association
Koh (2022) [59]	Non-myosteatosis compared to myosteatosis group: HR 0.648; 95% CI 0.486–0.865; p = 0.003	-
Kemper (2023) [19]	+ 10 HU increase is associated with lower HR of death 0.63, 0.49–0.81, p < 0.001	-	-
Chen (2022) [42]	HR 1.489, 95% CI 1.147–1.932, p = 0.003	-	-
Lee (2023) [13]	-	-	Non-myosteatosis compared to myosteatosis group: HR 0.543, 0.392–0.753, p < 0.001
Okugawa (2018) [38]	No significant association	No significant association	No significant association
Malietzis (2016) [50]	No significant association	-	No significant association
Cho (2024) [60]	Age group 51–74: non-myosteatosis compared to myosteatosis group HR 0.63 (0.41 − 0.96), p = 0.033 Age ≥ 75: no significant association	-	-
Kim (2023) [61]	HR 1.90, 95% CI 1.84–3.04, p = 0.008	-	HR 1.89, 95% CI 1.25–2.86, p = 0.003
Yokoi (2022) [65]	-	-	HR 2.49, 95% CI 1.21–4.95, p = 0.01
Lee (2023) [13]	HR 1.55; 95% CI, 1.32–1.82, p < 0.05	-	-
Ferreira (2024) [31]	Stage I–III: HR adjusted 2.78, 95% CI 1.26–6.15, p = 0.012	-	-
Zhou (2024) [63]	No significant association	-	No significant association
Colon cancer
Chung (2020) [27]	No significant association	-	-
Xiao (2020) [25]	HR 1.44, 95% CI 1.12–1.85, p < 0.05	-	-
Kim (2023) [15]	HR 1.90, 95% CI 1.84–3.04, p = 0.008	-	HR 1.89, 95% CI 1.25–1.86, p = 0.003
Golder (2022) [62]	Stage II: (3-yr OS: myosteatosis 85% vs NM 96%, p < 0.001) Stage III: 3-yr OS myosteatosis 75% vs NM 76%, p = 0.034)	-	-
Pozzuto (2021) [22]	SMD: non-myosteatosis compared to myosteatosis group HR 0.91, 95% CI 0.87–0.95, p < 0.001 IMAT: HR 1.06, 95% CI 1.03–1.09, p < 0.001	-	SMD: Non-myosteatosis group compared to myosteatosis group HR 0.95, 95% CI 0.91–0.98, p = 0.003 IMAT: HR 1.03, 95% CI 1.01–1.06, p = 0.011
Rectal cancer
vanRees (2021)	No significant association	-	-
Pozzuto (2021) [22]	No significant association	-	No significant association
Giani (2022) [28]	-	-	No significant association
Wei (2024) [66]	High SMD: HR 0.24, 95% CI 0.06–0.86, p = 0.029; high IMAT volume: HR 0.13, 95% CI 0.03–0.59, p = 0.008; high IMAT density: HR 0.26, 95% CI 0.10–0.68, p = 0.006	-	High SMD: HR 0.45, 95% CI 0.2–0.99, p = 0.048

HR hazard ratio, CI confidence interval, HU Hounsfield unit

GRADE summary of evidence certainty

The quality of included studies was assessed using the Newcastle–Ottawa Scale (NOS) for cohort studies [12]. Only 16 of 54 studies were found to be of “good” quality, as assessed by the Newcastle Ottawa rating scale and Agency for Healthcare Research and Quality (AHRQ) standards (Table 4). Many studies were deemed to be at high risk of publication bias due to limited adjustment for potential confounders, insufficient follow-up periods for assessment of survival and a failure to report the number of participants lost to follow-up or the reasons for their loss. Overall, the certainty of evidence regarding the prognostic impact of myosteatosis in colorectal cancer is low to moderate. The evidence supporting an association between myosteatosis and overall survival (OS) is moderate-certainty, downgraded primarily for risk of bias due to the predominance of retrospective designs, incomplete adjustment for key confounders and incomplete information on follow-up data. Findings for colon cancer are more consistent and therefore of moderate certainty, whereas evidence for rectal cancer is inconsistent and of low certainty.

Table 4.

Assessment of study quality based on Newcastle–Ottawa scale

Study	Selection	Comparability	Outcome	Overall quality
Giani (2020) [28]	4	0	0	Poor
Carcamo (2021) [30]	4	1	2	Good
VanVugt (2018) [18]	4	2	3	Good
Chung (2020) [27]	4	1	0	Poor
Tankel (2020) [26]	4	1	1	Poor
Hopkins (2019) [53]	4	2	2	Good
Kusunoki (2021) [39]	4	2	1	Poor
Aro (2020) [45]	4	1	2	Good
Sueda (2018) [54]	4	2	1	Poor
Berkel (2019) [44]	4	2	1	Poor
McSorley (2018) [55]	3	1	1	Poor
Xiao (2023) [49]	4	1	1	Poor
Souwer (2020) [41]	4	2	1	Poor
vanBaar (2018)	4	2	1	Poor
McGovern (2023) [46]	4	0	1	Poor
Lee (2023) [13]	4	1	2	Good
Renman (2023) [64]	4	1	2	Good
Golder (2022) [62]	4	1	2	Good
Bimurzayeva (2024) [24]	3	1	2	Good
Looijaard (2020) [23]	4	1	2	Good
Pozzuto (2021) [22]	4	1	1	Poor
Schaffler-Schaden (2020) [21]	3	1	2	Good
Pekarova (2021) [33]	3	0	0	Poor
Blackwell (2023) [20]	4	1	1	Poor
Boer (2016) [43]	4	2	1	Poor
vanderKroft (2018)	3	2	1	Poor
Dolan (2019) [57]	3	0	1	Poor
Aro (2022) [45]	4	1	2	Good
Koh (2022) [59]	3	0	2	Poor
Benedek (2021) [34]	3	0	1	Poor
Kemper (2023) [19]	4	2	1	Poor
Chen (2022) [42]	4	0	1	Poor
vanRees (2021)	3	1	1	Poor
Lee (2023) [13]	3	1	0	Poor
Almasaudi (2019) [48]	3	1	0	Poor
Margadant (2016) [40]	4	1	1	Poor
Okugawa (2018) [38]	3	1	1	Poor
Xiao (2020) [49]	4	2	2	Good
Kim (2023) [61]	3	1	2	Good
Malietzis (2016) [50]	4	1	1	Poor
Cuijpers (2022) [16]	2	1	1	Poor
Cho (2024) [60]	3	1	0	Poor
Martin (2018) [51]	4	1	1	Poor
Herrod (2019) [35]	4	1	1	Poor
Pacquelet (2022) [17]	4	1	1	Poor
Kim (2023) [15]	4	1	1	Poor
Yokoi (2022) [65]	4	1	1	Poor
Giani (2022) [28]	4	1	2	Good
Ferreira (2024) [31]	4	1	1	Poor
Gao (2025) [36]	4	1	2	Good
Kotek (2024) [32]	4	0	2	Poor
Parnasa (2025) [37]	3	1	2	Good
Zhou (2024) [63]	4	1	1	Poor
Wei (2024) [66]	4	1	1	Poor

Thresholds for converting Newcastle Ottawa Scale to AHRQ standards

• Good quality: 3 or 4 stars in selection domain AND 1 or 2 stars in comparability domain AND 2 or 3 stars in outcome/exposure domain

• Fair quality: 2 stars in selection domain AND 1 or 2 stars in comparability domain AND 2 or 3 stars in outcome/exposure domain

• Poor quality: 0 or 1 star in selection domain OR 0 stars in comparability domain OR 0 or 1 stars in outcome/exposure domain

Evidence for the association between myosteatosis and severe postoperative complications (Clavien-Dindo ≥ III) is low to moderate certainty. Although several higher-quality studies show a significant association—particularly in colon cancer—substantial heterogeneity in myosteatosis definitions, measurement techniques and outcome reporting warrants downgrading for inconsistency and indirectness.

Across all outcomes, the certainty of evidence is further limited by substantial heterogeneity in CT acquisition protocols, diagnostic thresholds for myosteatosis, and inadequate adjustment for clinical and lifestyle confounders.

Discussion

This review synthesises 54 observational studies evaluating the prognostic and postoperative impact of myosteatosis in colorectal cancer (CRC). In 2023, Chang et al. conducted a systematic review and meta-analysis of 10 studies, finding a significant association between myosteatosis in stage I–IV colorectal cancer and OS (HR 1.52), CSS (HR 1.67), and RFS (HR 1.89) [70]. In contrast to these outcomes, we found that when focusing on populations with non-metastatic disease only, the evidence is heterogenous and nuanced.

Our research adds the notable observation that the prognostic effect of myosteatosis in colon and rectal cancer appears to vary. A significant impact of myosteatosis was seen in CRC; however, subgroup analysis revealed more consistent observations of reduced survival in colon cancer compared to rectal cancer, in both OS and DFS. There was a paucity of cancer subtype-specific analysis, especially in CSS and DFS. Myosteatosis in colon cancer was observed to impact OS with HR up to 1.90 in one study of good methodological reliability [13]. Notably, treatment regimens for colon cancer (CC) and rectal cancer (RC) differ significantly, with neoadjuvant chemotherapy efficacy well-established in rectal cancer and potentially acting as a confounding variable [71]. Further studies examining the two cancer subtypes separately will be required to clarify the impact of myosteatosis as a prognosticator.

The evidence to support an association between myosteatosis and postoperative outcomes in CRC was inconsistent. Studies examining the outcome of overall postoperative outcomes were most numerous; however, the evidence was not suggestive of a clear link. As identified by an earlier systematic review by van Helsdingen et al. this may be attributed to the heterogeneity in the types of postoperative complications that are found reported among these studies [4]. To build upon this, we have attempted to narrow the outcome of postoperative complications to include three more specific indicators of postoperative course.

The studies examining severe (Clavien-Dindo III and above) complications demonstrated a consistent correlation, with a particularly consistent correlation observed in colon cancer. In colon cancer, Xiao et al. reported a hazard ratio of 2.42 for severe complications in patients with myosteatosis [47]. Given the morbidity and cost impacts of postoperative complications graded as Clavien-Dindo 3 and above, the increased risk of such magnitude may be relevant in surgical decision-making [72]. However, there is a need for further research to validate this finding. The evidence surrounding the complications of anastomotic leak was inconsistent in mixed CRC populations, but significant findings emerged in colon cancer in sub-type analysis. The evidence regarding LoS was ambiguous in its clinical significance, with magnitudes of differences in mean LoS reported as 1 day.

The mechanism behind the worsened OS outcome and predisposition for severe postoperative complications has not been fully elucidated. An interplay between systemic inflammation, nutritional status, and body composition has been proposed, suggesting that myosteatosis may reflect a broader catabolic state that leads to metabolic dysfunction and cancer cachexia [73, 74]. Myosteatosis is associated with worsened physiological fitness, with studies demonstrating poorer outcomes in cardiac and respiratory disease, suggesting poorer physiological recovery from surgery [75]. However, further research into the exact cardio-respiratory complications experienced by patients of colon and rectal cancer surgeries is required to clarify this potential association.

This review has focussed on a qualitative approach, partly due to the wide variety of measures of myosteatosis found and the ongoing lack of consensus on a clear definition for this condition. Both Martin and Xiao’s definitions were widely used in studies from Europe and Asia. IMAT was used within the Netherlands and Brazil, while IMAC, based on the multifidus muscle, was evaluated in Japan. In general, the evidence to support the efficacy of SMD, using TAMA for measures of survival and psoas muscle for measures of postoperative complications, was most frequent in the literature. Some emerging evidence to support the use of IMAT and other volume-related measures was observed, but this was less abundant.

In addition to this, studies frequently analysed colon and rectal cancer populations together in analyses of postoperative outcomes and survival, despite major differences in their therapeutic approach. Furthermore, similar to an earlier review, this review identified heterogeneity in the statistical handling of outcomes [4]. Studies used a variety of approaches to statistical analysis of common outcomes, including multivariable/multivariate analyses involving adjusting factors selected through clinical or statistical approaches. Finally, this review identified a high risk of bias in many studies in this review, predominantly due to inadequate description and inclusion of confounding factors in multivariate analysis. We propose that multivariate analyses should consider the common confounding variables: age, sex, tumour stage and location, BMI, delivery of neoadjuvant or adjuvant therapy and surgical approach at a minimum, with lifestyle factors such as smoking also being important [76, 77]. This would improve the clinical applicability and reliability of the results. Furthermore, for survival analyses, we recommend for follow-up periods to surpass 5 years, the standard duration of follow-up and benchmark for achieving cure of cancer. Follow-up durations and loss-to-follow-up information should be clearly stipulated to improve reliability.

The majority of studies elected to utilise distinct cut-offs of radiological density to determine myosteatosis. However, current literature suggests that the radiological attenuation of muscles differs by age, gender, race and is specific to certain muscles [78, 79]. The heterogeneity observed in the results may be attributed to the inadequate consideration of these factors in the diagnosis of myosteatosis. Supporting this, in Cho et al.’s subgroup analysis of CRC patients, a significant association between myosteatosis and overall survival was only found in the below 75 age group, suggesting that the definition of myosteatosis should vary with age [60]. Given the difficulties associated with finding a definitive cut-off for these measurements, it has also been suggested that myosteatosis measurement should be continuous; although our review has not found clear evidence to suggest its efficacy, a relative paucity in studies utilising this approach underlines a need for further research.

We found that less than half of the studies specified the use of intravenous contrast and the specific parameters used to capture the CT images. CT attenuation of muscle is sensitive to the use of contrast and also to the exact vascular phase at which these images are taken, necessitating the creation of clear guidelines for scan acquisition and definition of the condition [80]. Future research should focus upon delineating clear descriptions of the methods used for scan acquisition to enable a clear definition to be synthesised.

Limitations

The primary limitation of this review is the heterogeneity among the included studies for both the methods used to assess body composition and the definitions applied. Therefore, a meta-analysis was not performed in this study. We also acknowledge that our literature search was limited to three databases and to studies in the English language. Furthermore, with many studies originating from Asia and Europe, the geological generalizability of the findings may be limited. A large proportion of the included studies were retrospective, which can limit the ability to establish definitive causal relationships between myosteatosis and outcomes.

Future directions

Future research should prioritise standardisation of CT acquisition protocols, measurement techniques and diagnostic thresholds, while stratifying analyses by colon and rectal cancer given their distinct surgical and oncological treatment pathways. Large, prospective, multicentre cohorts with robust adjustment for confounders—including lifestyle factors and neoadjuvant and adjuvant therapies—are required to clarify prognostic significance.

Conclusion

A review of 54 studies conducted on the prognostic effect of myosteatosis on postoperative course and survival in colorectal cancer was conducted. Myosteatosis has a negative association with OS, with consistent effects observed in colon cancer. There is a negative association with severe postoperative outcomes, particularly in colon cancer.

These results are limited by the risk of bias and heterogeneity in measurement methods and statistical analysis. Future work is required to clarify the scan acquisition protocol and diagnostic criteria.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

CH and JZ collected and curated data. CH and JZ drafted the original manuscript. RK, MW and JY critically reviewed the manuscript for intellectual information. All authors edited the manuscript. JY was responsible for conception and design. JY supervised the study. All authors read and approved the final version of the manuscript. CH and JZ contributed equally and are equal first authors.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

All data used in this study were obtained from previously published articles identified through the systematic review process. The datasets generated and analysed during the current study are available within the article and its supplementary materials. No additional raw data were generated or analysed beyond those reported.

Declarations

Conflict of interests

The authors declare no competing interests.