Sarcopenia as a novel prognostic factor in the patients of primary localized gastrointestinal stromal tumor

Background : Sarcopenia predicts poor prognosis of a variety of gastrointestinal malignancies. However, there is a lack of study on the association between skeletal muscle index (SMI) and the prognosis of gastrointestinal stromal tumor (GIST). The aim of this study is to develop a novel nomogram based on sarcopenia for GIST patients to predict overall survival (OS) Methods : SMI was measured by computed tomography scan of 107 patients who underwent resection for primary localized gastrointestinal stromal tumor (GIST). Sarcopenia was dened by cutoff values for SMI as 40.1cm 2 /m 2 and 39.8 cm 2 /m 2 using optimum stratication for males and females respectively. Factors were included in the nomogram were specied by univariate and multiple Cox proportional hazard analysis. Concordance index (C-index) and calibration curves were conducted to measure the discrimination and accuracy of the nomogram. The utility of the nomogram was assessed by the decision curve analysis (DCA). Results: Twenty-eight (26.2%) of 107 patients were sarcopenic. Sarcopenia was correlated signicantly with BMI, albumin, female sex, resection style, mitotic index, rupture status, survival. Sarcopenia was signicantly related to decreased overall survival (p=0.003).The nomogram including sarcopenia status, resection style and mitotic index had an excellent discrimination with C-index 0.794. The calibration curves represented a good accordance between the actual observation and nomogram prediction for overall survival. Decision curve analysis illustrated that the nomogram was helpful in clinic . Conclusions: we developed a nomogram based on sarcopenia to predict overall survival after resection of GISTs which is an effective and favorable prognostication tool.


Background
Gastrointestinal stromal tumor (GIST) is the most common mesenchymal neoplasm occurring from the gastrointestinal tract [1]. Worldwide, the prevalence of GIST is 4.3-22 per million per year, with approximately mid 60s of age and equal gender distribution [2]. Before millennium, traditional remedy, for instance, chemotherapy and radiation were not effective on GIST. Identi cation of the driver mutations in c-KIT and platelet-derived growth factor receptor α (PDGRF-α) further led to the successful targeted therapy of tyrosine kinase inhibitor (TKI)-imatinib [3]. Despite advances in treatment modalities, radical resection with TKI treatment remains mainstay for GIST. Especially, the GIST patients with the intermediate and high risk are recommended for the TKI treatment postoperatively. Risk strati cation might differentiate with patients who need TKI treatment and those who do not. Therefore, a more precise and re ned strati cation system is truly needed to manage postoperative therapeutics.
In 2002, Fletcher [4] rstly developed the predicting system for GIST, presently described as the National Institute of Health (NIH) criteria comprising two factors (mitotic index and tumor size). In 2006, the Armed Forces Institute of Pathology (AFIP) criteria [5] added tumor location to NIH criteria. In 2008, the NIH risk strati cation system compared to AFIP was modi ed to add tumor rupture status. This modi ed NIH criterion has been widely accepted because it is more easier to apply than the AFIP criteria, and subsequent Memorial Sloan-Kettering Cancer Center (MSKCC) nomogram [6]in 2009 and Joensuu's contour maps [7]in 2012. All the GIST recruited in this cohort were con rmed as the intermediate and high risk category based on modi ed NIH classi cations (tumor size, mitotic rate, location, and rupture).
Sarcopenia is an age-related disorder de ned by accelerating depletion in muscle mass and decline in strength and physical performance. Recently, sarcopenia has been regarded as the predictor of poorer disease free survival (DFS) or overall survival (OS) in many malignances [8][9][10]. Sarcopenia frequently occurs in the disease of chronic disorder such as cancer. In addition, aging could easily induce muscle atrophy [11] , and GIST likely occur in the mid 60's [2]. So, we raised a hypothesis that there might be a relationship between sarcopenia and outcome of GIST.
There are various methods to evaluate the sarcopenia, such as muscle mass, muscle strength and physical performance. Measured value of muscular mass area by a cross-sectional computed tomography (CT) image at the level of the third lumbar vertebra (L3) , namely skeletal muscle index (SMI) are generally accepted [12]. Due to SMI based on the preoperative imaging, sarcopenia might be a potential biomarker in uencing the prognosis of GIST. However, currently, there were few reports on the role of sarcopenia in patients with GIST.

Study design
This study was a single-center retrospective analysis. We evaluated 107 patients pathologically proven to be GIST with intermediate or high risk category who underwent surgical resection in the department of general surgery, the rst hospital of China medical university from February 2013 to February 2019. All the patients were followed up postoperatively by telephone interview, outpatient visits and China's native app-WeChat and all surviving patients were followed up for 6 years. Patient status was due to the time of last follow-up as follows: alive and dead. Survival time was determined as starting from the date of rst operation until end of follow-up due to either death or end of data collection. Only "dead" was considered an event in the analysis of overall survival (OS). This research was approved by Ethics committee of the rst hospital of China medical university.

CT valuation
Every patient of CT-image valuation was performed using ImageJ software. . Skeletal muscle volume was evaluated to use the CT image for all the eligible patients. A transverse CT examination at the third lumbar vertebra (L3) of the single 5mm-slice was assessed. The muscles in the L3 region-including internal and external obliques, transverse and rectus abdominus, psoas, quadratus lumborum, and erector spinae were analyzed.
L3 skeletal muscle index (L3 SMI, cm 2 /m 2 ) is calculated as follows: cross-sectional areas measurements of the muscle (cm 2 ) at the L3 divided by the height.
Our own sex-speci c cut-off values for the L3 SMI were established at which the survival difference was most signi cant, we used optimum strati cation to de ne the SMI cut-off value 39.8 cm2/m2 for women and 40.1 cm2/m2 for men. SMI below the de ned cutoff value was described as sarcopenia. The method of optimum strati cation has been previously described in literature [13] to separate sarcopenic patients and non-sarcopenic patients.

Statistical analysis
The characteristics of the 2 groups (sarcopenia and non-sarcopenia) were compared. Optimum strati cation is applied to nd the most signi cant p value by use of the log-rank χ² statistic to de ne the sex-speci c cut-offs associated with overall survival in our cohort. Optimum strati cation solves the threshold value of the continuous covariate (skeletal muscle index) which, based on log-rank statistics, best separates sarcopenic patients with non-sarcopenic patients due to end point (death). Univariable analysis was performed using c 2 test, the Mann-Whitney U test for categorical variables and independentsample t test for continuous variables. The Kaplan-Meier method was used to estimate OS. Variables with p < 0.05 in univariable analysis were included in multivariable analysis. A Cox proportional-hazards regression analysis was performed. The predictive performance of nomogram related to survival was assessed by C-index. The utility of the nomogram was evaluated by decision curve analysis (DCA). P value of less than 0.05 was considered signi cant. All statistical analyses were performed with the Statistical Package for Social Sciences (SPSS) 24.0 for Windows (Chicago, Illinois) and R software (version 3·5·0).

clinical parameters of GIST patients
Among 107 enrolled patients, 28 (26.2%) were classi ed as the sarcopenia group and 79 (73.4%) into the non-sarcopenia group. The clinical characteristics (Table 1) between the two groups were compared. The mean value of BMI and serum albumin in the non-sarcopenia group was higher than in the sarcopenia group (P = 0.001; P=0.007). Signi cant difference in resection style (p=0.034), mitotic index(p=0.02), rupture status(p=0.007), survival status(p=0.001) were found. Age, hemoglobin, gender, tumor site, tumor size and pathological type were not observed to differ signi cantly in two groups.
We also explored the relationship between sarcopenia and the degree of postoperative complications according to the Clavien-Dindo classi cation (table 2). Complications occurred in 46 cases. The majority of complications were concentrated in grade and grade , which accounted for 87% (40/46). Five patients were performed by second operation or interventional therapy due to anastomotic leakage and abdominal infection, which were classi ed as complication of grade III. One patient occurred heart failure of grade IV complications. None died in the perioperative period. The incidence of complications differed signi cantly in the two groups (p=0.001).
By univariate analysis, sarcopenia (p=0.003), resection style (p=0.001) and mitotic index (p=0.005) were signi cant predictors of OS. They were also further conducted in the multivariate analysis. sarcopenia (p=0.016), resection style (p=0.036) and mitotic index (p=0.044) were still identi ed as the independent prognostic factor for OS in GIST patients.

Clinical use of DCA curve analysis
The decision curve analysis (shown in the FIG 4) elucidated that the nomogram was feasible to make clinical valuable decision.

Discussion
Our research showed a comprehensive analysis from clinicopathological data and CT scans to explore the survival outcomes in the Chinese population with GIST following surgery. Preoperative sarcopenia de ned by SMI was associated with poor survival. CT has a high degree of validity in assessing body composition and is regarded as the gold standard method for estimating muscle [10,14,15].To the best of our knowledge, this is the rst research indicating that impaired effects of sarcopenia for the survival in patients with GIST worldwide.
Sarcopenia is a syndrome affecting innumerable people with cancers and is independent predictor of detrimental outcomes such as physical disability, poor quality of life, and reduced survival [16]. Sarcopenia is considered a signi cant constituent of cancer cachexia syndrome, and the pathophysiology is involved with the systemic in ammation, including anabolic and catabolic pathways [17]. Proin ammatory cytokine such as cytokines tumor necrosis factor (TNF-α), interleukins (ILs) regulate the anabolic pathways via lipid mobilization and protein catabolism [18]. These cytokines could also inhibit myocyte differentiation and mediate the atrophy in skeletal muscle through SMAD2/3 signaling [19]. What's worse, the loss of skeletal muscle could further cause local in ammation, contributing to increasing systemic in ammation [17]. Some studies demonstrated association between skeletal muscle mass and systemic in ammatory markers in many cancers [20][21][22][23]. These correlations provide the solid basis of in ammation mechanism responsible for sarcopenia.
However, the clinical de nition of SMI remains inconclusive, the most widely used de nitions were de ned by Prado [13] in the western people. However, these de nitions might not be applicable to Chinese GIST patients because BMI and physique differs greatly between eastern and western populations. Therefore, we used optimum strati cation analysis to de ne the SMI cutoff. The incidence of sarcopenia with GIST patients was 26.2 % (28/107) in our study is consistent with the previous study [24] with sarcopenic morbidity of 38.7% in the GIST patients.
Sarcopenic patients represented lower BMI and serum albumin. These two factors were associated with sarcopenia but not with the survival. Sarcopenia is considered to be a better predictive tool than BMI and albumin for survival. BMI is not available to evaluate the body composition. Sarcopenic obesity may mislead clinical decision concerning nutritional treatment because sarcopenia could exist in the normal and overweight patients with elevated BMI [20]. Albumin is a negative acute-phase protein that decreases in concentration with ongoing systemic in ammation, poor health, and malnutrition which lead to the decreased skeletal muscle mass [23], hence, lower albumin might be associated with low-SMI value to re ect the sarcopenic condition. This study is conformable with previous ndings in other cancers [25,26].
The high e cient predictive tool is indispensable for GIST patients. At present, a few in uential predictive models for the prognosis of GIST have been designed. Miettinen [5]proposed that tumor size, mitotic rate, and tumor site can accurately predict the risk of GIST patients. Gold developed a nomogram for the recurrence free survival of patients with GIST patients after complete resections [6]. This study established a new nomogram including resection style, mitotic index and sarcopenia. To prove the clinical validity, we evaluated whether the nomogram-assisted decisions would be bene cial to patient outcomes or not. The novel method of decision curve analysis demonstrated that if the threshold probability of a patient or doctor is 40%, then probably 7.5 persons would bene t without detriment of others. Our nomogram also represented the reliable performance with high c-index of 0.794. Our nomogram did not include other important clinical factors, for instance, the tumor size, tumor site and presence of rupture which have been reported to be correlated with the mortality of GIST patients [27][28][29]. The possible reason might be some patient records affecting the data bias was excluded for insu cient clinical data. In the future study, we need the more clinical samples and validation set to prove our nomogram.
To reinforce the precision of tumor biological behavior prediction by only depending on resection style and mitotic counts, we integrated the new parameter sarcopenia by measurement of SMI into our nomogram. Preoperative and postoperative SMI are of equal importance. An interesting study [24] showed 63.6% of initially sarcopenic GIST patients became non-sarcopenic after 6 months of imatinib. This reversal might be explained by the drug's anti-tumor activity. Hence, clinicians should persist in nutritional guidance for the whole management of GIST patients. Meanwhile, patients are encouraged to exercise and receive appropriate nutritional treatment for muscle protein synthesis against sarcopenia [11,30].
This study had several limitations. First, according to the European Working Group on Sarcopenia in Older People (EWGSOP), sarcopenia should be measured by the parameters of muscle mass, muscle strength and physical performance [31]. We regarded only muscle mass as the de nition of sarcopenia due to the retrospective design, and prospective study is further needed including more assessment tools for sarcopenia. Second, merely 107 patients were enrolled due to insu cient available clinical data and low incidence of GIST. Third, this was a single-institution study of small sample, and whether the results are feasible for other patient sets which needs further internal and external validation. Nevertheless, to best of our acknowledgement, this is the rst study to establish a predicting model based on the preoperative sarcopenia of GISTs patients. More variables, for instance, in ammation index and gene detection could be incorporated in future.

Conclusion
In summary, we conducted a comprehensive analysis of preoperative sarcopenia associated with survival prognosis and established a new nomogram accurately predicting 3-and 5-survival for GIST patients.
Thus, this study might be helpful for the physician to make better clinical evaluation.

List Of Abbreviations
Gastrointestinal stromal tumor, GIST; Overall survival, OS Declarations Ethics approval and consent to participate The study was approved by the ethics committee of the rst hospital of China medical university (no. 201711030). Patient consent was not required to review their medical records by the ethics committee of the rst hospital of China medical university of its retrospective design, and exemption from informed consent did not adversely affect the health and rights of subjects. This study kept con dentiality of patient data and strictly complied with the Declaration of Helsinki and its later amendments or comparable ethical standards.

Consent for publication
Not applicable Availability of data and materials The datasets during and/or analyzed during the current study available from the corresponding author on reasonable request.

Competing interests
We declared no competing interests Funding This work was supported by China Medical University Foundation for young key teachers (No. 3110118149), Natural science foundation of Liaoning province (No. 2019-BS-282). The funders did not have any in uence on any aspects of the study (i.e., design, data collection, analyses, interpretation, or writing the manuscript).
Authors' contributions HS wrote the article including: data analysis, grammar editing, manuscript.MD conceptualize the whole idea of text, and edit the manuscript. All authors have read and approved the manuscript Non-sarcopenia (n=79)     The Decision Curves Analysis curve of the predictive nomogram including three factors (sarcopenia status, resection style, mitotic index). The horizontal axis represents the threshold value, which is the where the expected bene t of treatment was equal to the expected bene t of avoiding treatment and the vertical axis represents adding up the true positive results and subtracting the false positive results. The nomogram (red line) has the high value due to the larger net bene t,