Influence of nutritional status and body composition on postoperative events and outcome in patients treated for primary localized retroperitoneal sarcoma

Abstract Background Retroperitoneal sarcomas (RPS) are rare tumours of mesenchymal origin, commonly presented as a large tumour mass at time of diagnosis. We investigated the impact of body composition on outcome in patients operated on for primary localized RPS. Patients and methods We retrospectively analysed data for all patients operated on for primary RPS at our institution between 1999 and 2020. Preoperative skeletal muscle area (SMA), visceral and subcutaneous adipose tissue area (VAT and SAT) and muscle radiation attenuation (MRA) were calculated using computed tomography scans at the level of third lumbar vertebra. European Working Group on Sarcopenia in Older People (EWGSOP2) criteria were applied to define myopenia. Using maximum log-rank statistic method we determined the optimal cut-off values of body composition parameters. Myosteatosis was defined based on determined MRA cut-offs. Results In total 58 patient were eligible for the study. With a median follow-up of 116 months, the estimated 5-year overall survival (OS) and local-recurrence free survival (LRFS) were 66.8% and 77.6%, respectively. Patients with myopenia had significantly lower 5-year OS compared to non-myopenic (p = 0.009). Skeletal muscle index and subcutaneous adipose tissue index predicted LRFS on univariate analysis (p = 0.052 and p = 0.039, respectively). In multivariate analysis high visceral-to-subcutaneous adipose tissue area ratio (VSR) independently predicted higher postoperative complication rate (89.2% vs. 10.8%, p = 0.008). Myosteatosis was associated with higher postoperative morbidity. Conclusions Myopenia affected survival, but not postoperative outcome in RPS. Visceral obesity, VSR (> 0.26) and myosteatosis were associated with higher postoperative morbidity. VSR was better prognostic factor than VAT in RPS.


Introduction
2][3] Most patients develop large tumour mass before diagnosis is clinically confirmed.Imaging techniques, CT and MRI are primarily used in clinical evaluation of RPS. 1,4urgical resection with removal of all gross disease is the cornerstone of curative therapy and optimal results are achieved with en bloc resection at the time of primary presentation. 4,5The role of radiation therapy and chemotherapy in management of RPS is still under investigation.7][8] There are no currently available data supporting the use of routine neoadjuvant or adjuvant chemotherapy for these patients. 11][12][13] Institute of Oncology Ljubljana is the only referral sarcoma centre in Slovenia. 14ody composition changes are related to nutrition status and associated with perioperative outcome and influence management of surgical oncology patients. 157][18] Sarcopenia is a clinical syndrome in which involuntary loss of skeletal muscle mass and function is progressive and generalized, together or without increased fat mass. 19,20nother clinically important body composition abnormality, myosteatosis, is characterized by excess accumulation of adipose tissue within muscle, resulting in impaired muscle strength and physical ability, as well as increased frailty. 21,223][24][25][26][27] Visceral obesity (VO) is the fat accumulation in visceral adipose tissue and serves as a clinical marker for adiposopathy. 285][36][37] Also, another body composition abnormality, the new concept of sarcopenic obesity (SO), a combination of excess adiposity and sarcopenia, seems to have powerful negative prognostic impact in oncology treatment and is gaining increased attention in cancer research. 38,39Loss of muscle mass or myopenia is a critical determinant of sarcopenia.
1][42][43] There is a lack of literature data regarding the impact of body composition on postoperative and oncologic outcome in patients operated on for primary RPS.
The aim of our study is to investigate the impact of low muscle mass or myopenia, myosteatosis, visceral obesity and cancer cachexia on OS, local recurrence-free survival (LRFS) and postoperative morbidity in patients operated for primary localized RPS.Additionally, we aimed to investigate the predictive value of preoperative body composition parameters for OS, LRFS and postoperative morbidity.

Study design and population
Retrospective study was conducted on patients operated on for primary RPS at Department of Surgical Oncology at the Institute of Oncology Ljubljana between September 1999 and June 2020 (Figure 1).A total of 58 patients met the inclusion criteria, 24 females (41.4%) and 34 males (58.6%).The Slovenian National Medical Ethical Committee (decision number: 0120-530/2020/3), Institutional Review Board (ERID-KSOPKR-0081/2020) and Institutional Ethical Committee (ERIDEK-0079/2020) approved the study.Due to the retrospective nature of the study the need to obtain informed consent from participants was waived.

Clinical data collection
Patient's histories including anesthesiologic preoperative reports, operative reports, hospital records, and follow-up data were reviewed.Clinical and pathological data were collected (Figure 1).Postoperative complications were evaluated in accordance with Clavien-Dindo classification. 44umour features of interest were as follows: histopathological diagnosis, stage (according to American Joint Committee on Cancer [AJCC] 8th Edition), grade (according to National Federation of Centers for the Fight Against Cancer grading system) 45 and tumour size (largest diameter value).Resection quality was recorded as either complete (R0), incomplete (R1) with microscopic involvement of resection margins or macroscopic residual tumour (R2).

Body composition assessment
The assessment of body composition was conducted using images from CT scans taken within 30 days preoperatively at the level of the third lumbar vertebra using the "Automated Body Composition Analyzer using Computed tomography image Segmentation" (ABACS) software. 46This method uses predefined CT Hounsfield units (HU) values to recognize different tissues.The CT HU thresholds were 29 to 150 for skeletal muscles, 190 to 30 for subcutaneous adipose tissue, and 150 to 50 for visceral adipose tissue.The following body composition parameters were measured: total cross-sectional skeletal muscle area (SMA, cm 2 ), subcutaneous adipose tissue area (SAT, cm 2 ) and visceral adipose tissue area (VAT, cm 2 ).After normalization by patient's height (m 2 ), we used these parameters as lumbar skeletal muscle index (SMI, cm 2 /m 2 ), subcutaneous adipose tissue index (SATI, cm 2 /m 2 ), and visceral adipose tissue index (VATI, cm 2 /m 2 ).VSR was calculated by dividing VAT by SAT.To assess the muscle density and myosteatosis, skeletal muscle radiation attenuation (MRA) has also been recorded in HUs.All measurements were performed by experienced researcher, accredited for complex image analysis and segmentation techniques.Additionally, we used previously reported and validated equations to calculate appendicular skeletal muscle index (ASMI), lean body mass (LBM) and fat mass (FM) 26,29,43,47 : ASMI (kg/m 2 ) = 0.11 x SMI (cm 2 /m 2 ) + 1.17 LBM (kg) = 0.030 x Lean Tissue Area (cm 2 ) + 6.06 FM (kg) = 0.042 x Total Fat Area (cm 2 ) + 11.2 Based on a single abdominal CT image per patient, LBM and FM properly reflect appropriate dual-energy X-ray absorptiometry (DXA) derived whole-body fat-free mass (FFM) and whole-body fat mass (FM), respectively.

Assessment of myopenia, myopenic and visceral obesity, myosteatosis and cancer cachexia
Myopenia was defined based on the new recommendations of The European Working Group on Sarcopenia in Older People (EWGSOP2), as follows: SMI < 43 cm 2 /m 2 for men with BMI < 25, SMI < 53 cm 2 /m 2 for men with BMI ≥ 25, and SMI < 41 cm 2 /m 2 for women. 48uscle mass in patients with obesity was assessed according to The European Society for Clinical Nutrition and Metabolism (ESPEN) and the European Association for the Study of Obesity (EASO) consensus report for sarcopenic obesity. 39reviously reported diagnostic criteria for visceral obesity were applied: VAT > 163.8 cm 2 for men and VAT > 80.1 cm 2 for women. 29,39,47reoperative cancer cachexia was defined using Fearon et al. criteria. 49n order to establish optimal cut-off values for SMI, VATI, SATI, VSR and MRA which would best reflect our study cohort in relationship to defined outcome (maximum OS), we performed optimal stratification analysis based on maximally selected rank statistics using maxstat package implemented in R statistics. 50,513][54][55] The presence of myosteatosis was then confirmed based on established optimal threshold for MRA: < 35.88 HU in patients with a BMI ≥ 25 kg/m 2 and < 47.41 HU in those with a BMI < 25 kg/m 2 .

Survival and statistical analysis
Final survival follow-up time was set as last follow-up date in the study period or the event of death.OS was defined as time between the date of the operation and date of death from any cause or last follow-up.LRFS was defined as the time interval between operation date and date of first documented local progression, and instances involving deaths without evidence of disease and the occurrence of distant metastases considered as competing events.Survival curves were estimated using Kaplan-Meier method.Log-rank test, linear regression and Cox proportional hazard regression models were used to analyse the relationship between clinicopathological parameters  Albumin level (g/dL), C-reactive protein (mg/L), neutrophil-to-lymphocyte ratio (NLR), preoperative radiotherapy, tumour size (cm), and intraoperative blood loss (ml).Results were statistically significant if two-sided p value < 0.05 was achieved.R statistical software (version 4.2.1, R core Team) was used.
The results of optimal stratification analysis for finding cut-off values for SMI, VATI, SATI, VSR and MRA are presented in Table 3 and Supplementaly Figures 1−4.

Association between overall morbidity following surgery for primary RPS and body composition (A−F) and linear correlation analysis between VSR and VATI (G) and VSR and fat mass (H). VSR (F) and intraoperative blood loss
In univariate analysis of overall postoperative morbidity, the presence of myosteatosis (OR 5.05, p = 0.023), VO (OR 3.61, p = 0.047) and high VSR (OR 6.19, p = 0.008) were associated with signifi-cantly higher overall complication rate.Adjusted for other covariates in multivariate analysis, high VSR maintained significant impact (adjusted OR 5.05, p = 0.05).We omitted VO from multivariate analysis to avoid multicollinearity.Figure 3 (panels A-F) summarises our analysis of morbidity following surgery for primary RPS.

Discussion
Our study provided new insight into the association between preoperative body composition and postoperative and oncologic outcome in primary RPS patients.We focused on evaluation of the sig- nificance of preoperative nutrition status-related syndromes.Furthermore, we examined the predictive value of SMI and MRA, measures of muscular quantity and quality, as well as, VATI, SATI and VSR, measures of adiposity, for possible clinical use in preoperative clinical assessment of patients diagnosed with this rare malignancy.To address the lack of literature and inconsistency in used body composition cut points, we used maximally selected rank statistics to defined cohort -specific cut point.This method incorporated follow-up time and time-to-event outcomes, dividing the patients into two groups with the most significant statistics between each other in term of survival. 56n our cohort, both myopenia and VO were associated with poorer OS.Patients with myopenia had 5-year OS of 33.7%, compared to significantly higher 78.3% 5-year OS for non-myopenic patients (p = 0.009).SMI predicted LRFS on univariate analysis and lost prognostic value in multivariate analysis.
These findings are in line with knowledge that sarcopenic surgical oncology patients are at greater risk for poor operative outcome because of underlying muscle mass loss which is an integral component of sarcopenia and also facilitates the impairment of muscle function and physical performance. 57Therefore we used the cut-off values for diagnosis of myopenia which are the component of diagnostic criteria and tools that define and characterize sarcopenia in EWGSOP2 Revised European Consensus. 48Our optimal fitting method analysis for establishing the cut-off value for defining low SMI (used for comparative and descriptive purposes) resulted in slightly different cut-off values: SMI < 49.21 cm 2 /m 2 for males and females with BMI < 25, and SMI < 49.9 cm 2 /m 2 for males with BMI ≥ 25 and SMI < 50.64 cm 2 /m 2 for females with BMI ≥ 25.Both EWGSOP2 criteria for low SMI and our cut-off values were able to predict poor prognosis.It seems, that difference is generated because our cohort consisted of only patients with primary RPS with resectable disease rather than a heterogeneous cohort.We also found that SMI analysed as continuous variable was not able to predict poor outcome.This is another proof that in clinical practice SMI should be evaluated as body composition (myopenia) parameter defined with cut-off values below which the risk of poor prognosis is increased significantly, rather than discretional decrease. 580][61] Most of this data is founded on reports about patients operated on for gastrointestinal cancers.In our study cohort myosteatosis was also associated with greater overall complication rate (OR 5.05, 95% CI 1.39-24.41,p = 0.023) in univariate analysis, but it was not confirmed in multivariate analysis.Myosteatosis was not associated with OS, LRFS or postoperative outcome.However, recently a group of authors reported significant association between myosteatosis and major complication rate and OS in retroperitoneal and trunk soft tissue sarcoma. 29They used preoperative MRA as continuous variable to define myosteatosis, not providing any cut-off point for reference.We defined myosteatosis based on optimal cut point analysis for MRA, and determined cut-offs are comparable to most commonly used range of MRA cut-offs for myosteatosis. 62n order to evaluate obesity and the distribution of fat tissue, we calculated VAT, SAT and corresponding height-adjusted indexes VATI and SATI.We also considered BMI.Higher value of BMI (≥ 25) was not associated with oncologic or postoperative outcome.4][65] Stratified for myopenia, comparison of the subgroups revealed that body fat and VATI were significantly higher in non-myopenic patients (median 31.1 vs. 28.1%,p = 0.024 and 42.3 vs. 20.6 cm 2 /m 2 , p = 0.025, respectively) (Table 2).Further on, we used VAT to assess VO applying the ESPEN/EASO criteria. 39VO predicted poorer OS and higher postoperative complication rate.VAT alone had no impact on OS or postoperative outcome.Recent study on soft tissue sarcoma patients reported identical findings. 29We considered VSR into adiposity analysis defining subgroup of patients with normal and high VSR (> 0.26) based on optimal cut-off analysis.In multivariate analysis VSR was an independent predictor for overall complication rate following surgery.High VSR group experienced significantly more complications compared to normal VSR group (33 (89.2%) vs. 4 (10.8%),p = 0.008).These results are comparable with previous reports in which VSR was superior to VAT as independent risk factor for death and local recurrence. 34,35,58,66,67Linear regression analysis showed significant corelation between VSR and both VATI and fat mass (Figure 3 -panels G−H and Supplementaly Figure 5), confirming the importance of balance between visceral and subcutaneous adipose tissue.Recent studies demonstrated that predictive values of VSR for cardiovascular and metabolic disease incidence is superior to VAT. 34,36,37,68 However, to our knowledge, only a few studies investigated the impact of VSR and VAT on survival and postoperative outcome in patients operated on for primary RPS. 29Our study underlined that high VSR is not only superior to VAT but also to BMI in predicting poor oncologic and perioperative outcome.These findings suggest that VSR better estimates adipose tissue distribution and poses an additional difficulty for performing the surgery itself.High VSR is strong independent predictor for overall postoperative morbidity (multivariable-adjusted OR 5.05, p = 0.05).On the other hand, in the context of survival analysis, the multivariate regression model was not able to confirm the statistical significance of VSR (p = 0.068).This suggests that the impact of VSR on survival of RPS patients may be attenuated when considered alongside the broader set of predictors.One of the reasons may be the fact that, the presence of high VSR, as determined by specific gender-independent cut-off criteria, exhibited a statistically significant gender difference, with females having higher odds (OR = 4.5, p = 0.027) compared to males.Furthermore, we found a statistically significant difference in the distribution of SMI, between the two groups defined by VSR (high VSR vs. low VSR OR = 0.926, p = 0.047).Logistic regression model revealed significant association of SMI with a reduced odds of the specified outcome within the "high VSR" group.Based on our initial hypothesis that "high VSR" has a negative impact on survival, supported by univariate analysis, this implies that SMI (approximation of myopenia) may be a factor that mitigates the negative impact of "high VSR" on patient survival or serves as a positive influence, hence confounding the effect of VSR in multivariate settings.Further prospective studies need to be developed to confirm the importance of preoperative VSR for poor postoperative survival.In contrast to high VAT, low SATI, independently predicted poorer OS (adjusted HR 7.00, p = 0.057).][71] The multivariate analysis demonstrated that, when assessed as a continuous variable, albumin levels (HR 0.9, 95% CI 0.81−0.98,p = 0.019) and NLR (HR 1.4, 95% CI 1.06−1.75,p = 0.015) were independently associated with overall survival.This finding underscores the pivotal role of these inflammatory biomarkers in clinical practice and management of surgical oncology patients.3][74][75][76][77] Furthermore, our observation suggests that hypoalbuminemia identifies a high-risk cohort that may derive greater benefits from enhanced nutritional support preoperatively.The omission of descriptive statistical analysis for serum albumin and NLR, in term of patients' outcome, limits the depth of our data exploration.
Our study had some limitations.It was a single center report including relatively small number of patients which might influence the power of drown conclusions.Another weakness was the fact that we didn't assess all comorbidities in our analyses, as they were considered negligible in patients with soft tissue sarcoma.However, since our Institution is the only referral sarcoma center in Slovenia, having population of 2.1 million, our study cohort consisted of unique set of primary RPS patients eligible for curative surgery.We reported the most distinguishable, cohort -specific, cut points for CT measured body composition (muscle and adipose tissue) parameters in regard to long term prognosis.And finally, providing a unique and new insight into the association between preoperative body composition and postoperative and oncologic outcome in primary RPS patients was the main strength of the study.

Conclusions
Patients with primary RPS are in a great risk for nutritional disorders for number of reasons such as: requirement for demanding abdominal surgery in their management, long preclinical history and tumour size. 4,10,18,78,79In our study cohort there was a high prevalence of myopenia (32.8%) and visceral obesity (36.2%).Myopenia, cancer cachexia and low SATI were strongly associated with poor OS.High VSR was strong independent predictor for overall postoperative morbidity.Additional prospective studies are required to substantiate the role of preoperative VSR as independent prognostic factor for postoperative survival.Our findings suggest that clinical nutrition interventions towards improving visceral adiposity and myopenia may benefit surgical and oncologic outcome in primary RPS patients.

F
(E) independently predicted worse postoperative outcome.In multivariate analysis skeletal muscle index (SMA), lean body mass (LBM), subcutaneous adipose tissue area (SAT) and visceral adipose tissue area (VAT) were not associated with statistically higher overall postoperative morbidity (A−D).VATI = Visceral Adipose Tissue Index; VSR = Visceral-to-subcutaneous adipose tissue area ratio.ȓ Pearson = Pearson Correlation Coefficient; t Student = result of t-test for correlation postoperative morbidity, intrahospital length of stay (LOS) and major postoperative complications.

TABLE 1 .
Clinical characteristics of study population AJCC = The American Joint Committee on Cancer; ASA = American Society of Anesthesiologists classification; EWGSOP2 = The European Working Group on Sarcopenia in Older People; FNCLCC = Fédération Nationale des Centres de Lutte Contre Le Cancer Summary for continuous variables is presented as median (interquartile range) and the statistical test is Kruskal-Wallis/Mann-Whitney; a cut-off values displayed in Table3

TABLE 2 .
Comparison of clinical and body composition parameters between myopenic and non-myopenic patients (EGSWOP2 criteria) a Summary for continuous variables is median (interquartile range) and the statistical test is Kruskal-Wallis/Mann-Whitney; b Median (IQR); n (%); AJCC =The American Joint Committee on Cancer; ASA = American Society of Anesthesiologists; EGSWOP2 = The European Working Group on Sarcopenia in Older People; FNCLCC = Fédération Nationale des Centres de Lutte Contre Le Cancer; HU = Hounsfield units The demographic and clinical characteristics of the patients are provided in Table1.In the cohort, 34 (58.6%) were males and 24 (41.4%) were females.Median age at diagnosis was 61.0 (46.0−67.0).

TABLE 3 .
Results of optimal stratification analysis for body composition parameters

BMI, kg/m 2 Skeletal Muscle Index a , cm 2 /m 2 Visceral Adipose Tissue Index b , cm 2 /m 2 Subcutaneous Adipose Tissue Index b , cm 2 /m 2 Visceral to subcutaneous ratio c Muscle Radiation Attenuation d , HU
Hounsfield Unit; a adjusted for gender and BMI; b adjusted for gender only; c cut-off determined on the level of whole cohort, not stratified for BMI nor gender; d adjusted for BMI only.

TABLE 4 .
Summary of univariate and multivariate analysis of association between body composition and outcome following surgery for primary RPS

Clinico-pathological factor Length of hospital stay (> 10 days) Clavien-Dindo > IIIa Any complication (overall morbidity)
assessed by the European Working Group on Sarcopenia in Older People revised criteria from 2018; b defined as VSR > 0.26; c defined as SATI > 49.23 for males and SATI > 86.89 for females; d defined as VATI > 61.38 for males and VATI > 25.55 for females Only significant variables (p < 0.05) were included in multivariate analysis.