Prognostic impact of preoperative skeletal muscle change from diagnosis to surgery in patients with perihilar cholangiocarcinoma

Abstract Background Loss of skeletal muscle mass is a prognostic factor after surgery for gastrointestinal cancers. The treatment for perihilar cholangiocarcinoma (PHC) is a highly invasive surgery. Biliary drainage and portal vein embolization, which can prolong the preoperative waiting time (PWT), are often required before surgery. Assuming that the skeletal muscle mass can change during PWT, we investigated the clinical effect of skeletal muscle change on surgical outcomes of PHC. Methods We retrospectively reviewed the medical records of 89 patients who underwent curative surgery for PHC from January 2013 to December 2019. We defined the psoas muscle area (PMA) at the third lumbar vertebra as the skeletal muscle mass. The PMA just before surgery was divided by that at the time of diagnosis, and we defined it as the rate of change of PMA (CPMA). Patients were divided into two groups according to CPMA: wasting (n = 44, below the median CPMA) and no‐change (n = 45, above the median CPMA). Results The median PWT was 63 d, and CPMA was 96.1%. The median recurrence‐free survival and overall survival were significantly shorter in the wasting group than in the no‐change group (8.0 vs 33.2 mo, P = 0.001 and 14.2 vs 48.7 mo, P < 0.001, respectively). Multivariate analysis revealed that histological differentiation, R1 resection, lymph node metastasis, and preoperative skeletal muscle wasting were independent prognostic factors of PHC. Conclusion This study suggests that preoperative skeletal muscle wasting in patients with PHC has a negative effect on survival outcomes.


| INTRODUC TI ON
Perihilar cholangiocarcinoma (PHC) is one of the most aggressive malignancies, and its only potentially curative treatment is aggressive surgical resection, such as major hepatectomy with extrahepatic bile duct resection. [1][2][3] Before surgery, patients with PHC often require biliary drainage for obstructive jaundice and portal vein embolization to enlarge the remaining liver, and neoadjuvant chemotherapy can be adopted for advanced PHC. Thus, the preoperative waiting time (PWT) from diagnosis to surgery is often longer for PHC than it is for other malignancies. 4,5 Several studies have shown that sarcopenia, or the loss of skeletal muscle mass, is associated with poor short-and long-term outcomes in PHC. 6 The onset of sarcopenia can have multiple causes, including immobility, altered endocrine functions, and chronic inflammatory diseases (including malignancy). 7 Furthermore, rehabilitation and nutritional enhancement before surgery may affect sarcopenia status. 8,9 Most previous reports assessed patients' sarcopenia status only once before surgery. 10,11 However, in patients with PHC the sarcopenia status may change throughout the PWT due to cholangitis or other factors, and evaluation at a single point is considered insufficient to thoroughly assess the sarcopenia status of such patients.
Thus, research assessing preoperative skeletal muscle change in patients with PHC and its effects on short-and long-term outcomes is necessary.
We aimed to investigate the effect of preoperative skeletal muscle change during PWT on short-and long-term survival outcomes of patients who underwent PHC resection.

| Study design and patient population
We retrospectively reviewed the medical records of 111 patients with resectable PHC who underwent surgery at Keio University Hospital (Tokyo, Japan) between January 2013 and December 2019.
Of these, six patients who underwent palliative surgery or laparotomy due to distant metastasis or local cancer extension, three who underwent bile duct resection alone, six diagnosed with benign lesions, and seven diagnosed with other malignancies were excluded from the study. Finally, 89 patients who were pathologically diagnosed with PHC were included in the analysis ( Figure 1).
Study variables included patient characteristics, laboratory results, tumor characteristics, details of the surgery, complications, and overall survival and recurrence data. Inflammation-based prognostic scores, which indicate the nutritional status and predict F I G U R E 1 Flow chart of all patients who underwent surgery for suspected perihilar cholangiocarcinoma. HCC, hepatocellular carcinoma; cHCC-CC, combined hepatocellular and cholangiocarcinoma; GIST, gastrointestinal stromal tumor; NEC, neuroendocrine carcinoma prognosis, such as the prognostic nutritional index (PNI), controlling nutritional status (CONUT), Modified Glasgow Prognostic Score (mGPS), neutrophil-lymphocyte ratio (NLR), and plateletlymphocyte ratio (PLR), were calculated. [12][13][14][15][16] Laboratory results obtained from the patients 1-3 d before surgery were used to determine these prognostic scores. Pathologic examination of the tumors was performed according to the 8th edition of the Union for International Cancer Control tumor-node-metastasis classification system. 17 This study was approved by the Ethics Committee of the Keio University School of Medicine (Approval number: 20120443).

| Image analysis and definition of preoperative skeletal muscle wasting
All patients underwent abdominal/pelvic computed tomography (CT) at least twice: at the time of cancer diagnosis and just before surgery. The cross-sectional areas of the right and left psoas muscles (psoas muscle area: PMA) at the middle level of the third lumbar vertebra were measured using SYNAPSE VINCENT software (Fujifilm Co., Tokyo, Japan). The border of the psoas muscle was manually outlined and quantified using a Hounsfield unit threshold from −29 to +150. 18 The PMA value just before surgery was divided by that at the time of cancer diagnosis, and the resulting value was defined as the rate of change in PMA (CPMA) ( Figure 2). According to the median CPMA, patients were divided into the preoperative skeletal muscle wasting and no-change groups. The clinicopathological factors, as well as the short-and long-term outcomes of the two groups, were compared. The measured PMA was normalized according to height using the following equation: normalized PMA, defined as the psoas muscle mass index (PMI, cm 2 /m 2 ) = measured PMA (cm 2 )/height (m 2 ). The cutoff level of sarcopenia in PMI value was 6.36 cm 2 /m 2 for males and 3.92 cm 2 /m 2 for females. 19 Obesity was defined as body mass index (BMI) ≥ 25 kg 2 /m 2 . 20 Sarcopenic obesity was defined as patients with both sarcopenia and obesity.

| Preoperative management and surgical procedures
For patients who experienced jaundice before surgery, appropriate biliary drainage, either by endoscopic nasobiliary drainage, endoscopic biliary stent, or percutaneous transhepatic biliary drainage, was performed. If the remnant liver volume was less than 40%, portal vein embolization was performed 2-4 wk before surgery. We considered neoadjuvant chemotherapy for patients who could not undergo R0 resection due to severe local progression.
All surgical procedures were performed after total bilirubin concentrations in the serum decreased to <2 mg/dl. In most cases,

| Definitions of complications and follow-up
Postoperative complications were scored using the Clavien-Dindo classification system. 21 A major complication was defined as Clavien-Dindo grade III or higher within 30 d after surgery.
Posthepatectomy liver failure was defined as increased international normalized ratio value (or need for clotting factors to maintain a normal international normalized ratio) and increased plasma bilirubin F I G U R E 2 Measurement of the psoas muscle area (PMA) at the time of diagnosis and just before surgery. The bilateral psoas muscle cross-sectional computed tomography images at the middle-level of the third lumbar vertebra was measured using SYNAPSE VINCENT software. The border of the bilateral psoas muscle was manually outlined and quantified using Hounsfield unit thresholds (−29 to +150). The PMA just before surgery was divided by that at the time of diagnosis, and it was defined as the rate of change of PMA (CPMA) levels on or 5 d after the operation affecting clinical management (defined by the International Study Group of Liver Surgery as grade B or C). 22 Postoperative mortality was defined as death related to surgery. Clinical follow-up was performed routinely every 3 mo until the second y after surgery and every 6 mo until July 2021 or death.
Laboratory tests and follow-up CT scans were performed in the first 6 mo and later as necessary to detect recurrence. Adjuvant chemotherapy was considered for patients at pathological stage II or higher and for those with a performance status of 0 or 1.

| Statistical analysis
All analyses were performed using SPSS software (v. 26 Categorical data were reported as number and frequency (%), and continuous data were reported as median and interquartile range (IQR). The cutoff value of prognostic scores was determined as in previous reports. 16,23,24 The overall survival (OS) and recurrencefree survival (RFS) rates were calculated using the Kaplan-Meier method and compared using a log-rank test, except for patients who died related to surgery. OS was calculated from the date of surgery to the date of death or the most recent follow-up, and we defined the patients who died of other causes with no evidence of recurrence as censored cases. RFS was measured from the date of surgery to the date of disease recurrence. We performed univariate analyses using Cox proportional hazard regression models to select variables likely related to OS and RFS among all candidate clinicopathological or operation-related factors. Owing to the limited number of observed events, factors with a P-value <0.05 in the univariate analysis were used in the multivariate analysis with Cox proportional hazard regression models to determine the independent prognostic factors.
All results with two-tailed P-values <0.05 were considered statistically significant.

| Clinicopathological characteristics according to CPMA
The median time from diagnosis to surgery was 63 (IQR: 49-108) d. The median CPMA was 96.1% (IQR: 91.1%-101.2%). According to the median CPMA, we divided the 89 patients into two groups: 45 patients in the skeletal muscle no-change group and 44 patients in the skeletal muscle wasting group. Table 1 shows comparisons of clinicopathological characteristics between the two groups. We found that the PNI (P = 0.042) and CONUT (P = 0.020) were significantly worse in the skeletal muscle wasting group compared to the skeletal muscle no-change group. However, there was no significant difference in preoperative management, such as biliary drainage, percutaneous transhepatic portal vein embolization (PTPE), and neoadjuvant therapy, or the number of days from diagnosis to surgery between the two groups. The presence of sarcopenia or sarcopenic obesity at the time of diagnosis and just before surgery also did not differ significantly. Considering pathological factors, moderate/poor tumor differentiation was significantly common in the skeletal muscle wasting group (P = 0.024), while other factors displayed no significant difference.

| Operative factors and short-term results according to CPMA
In Table 2 Additionally, major complications and posthepatectomy liver failure tended to be more common in the preoperative skeletal muscle wasting group compared to the skeletal muscle no-change group.
Finally, mortalities were observed only in the preoperative skeletal muscle wasting group. Moreover, the OS was significantly shorter in the skeletal muscle wasting group according to the log-rank test (P < 0.001, Figure 3a).

| Effects of CPMA on long-term follow-up
After excluding three cases of mortality, the median RFS was 8.0 (95% CI 5.2-10.8) mo in the preoperative skeletal muscle wasting group and 33.2 (95% CI 13.8-52.6) mo in the no-change group.
The 1-and 3-y RFS rates were 36.6% and 15.5% in the preoperative skeletal muscle wasting group and 79.7% and 42.5% in the nochange group, respectively. Like OS, RFS was significantly shorter in the skeletal muscle wasting group according to the log-rank test (P = 0.001, Figure 3b).  skeletal muscle wasting was also an independent prognostic factor.

| Analysis of prognostic factors in PHC
To the best of our knowledge, this is the first report to focus on the clinical effect of preoperative skeletal muscle change in patients with resectable PHC.
Preoperative sarcopenia was reported to be a predictive factor of postoperative morbidities in several hepatobiliary cancers, including PHC. 6,25 Similarly, in the present study the preoperative skeletal muscle wasting group tended to have more major complications, including liver failure, and significantly longer postoperative hospital stays than the no-change group. Additionally, patients in the preoperative skeletal muscle wasting group had a lower PNI and higher CONUT score. This could reflect the malnutrition status in patients with preoperative skeletal muscle wasting, which could have a negative effect on postoperative recovery and tissue repair. 26 Other reasons for poorer perioperative outcomes in the preoperative skeletal muscle wasting group could be the higher rates of hepatopancreato-duodenectomy and intraoperative blood loss. Although the only observed pathological difference between the two groups was a histologic type, there is possibly more widespread bile duct cancer, which needed to combine pancreatoduodenectomy in the wasting group.
The prognosis after PHC resection depends on several tumorspecific factors, including lymph node metastasis, differentiation, Abbreviations: BMI, body mass index; CA19-9, carbohydrate antigen 19-9; CI, confidence interval; CONUT, controlling nutritional status; HR, hazard ratio; mGPS, modified Glasgow Prognostic Score; NLR, neutrophil-lymphocyte ratio; PLR, platelet-lymphocyte ratio; PNI, prognostic nutritional index. a Grade III or higher according to the Clavien-Dindo classification. and resection-based curability; however, most of these factors can only be determined after the operation. In clinical settings, prognostic factors that can be identified before the operation are more useful. Regarding patient-specific factors, several retrospective studies have shown that the survival rate of preoperative sarcopenic patients with PHC was significantly lower than that of nonsarcopenic patients. 6 In this study, preoperative skeletal muscle wasting was one of the significant independent prognostic factors, while other preoperative factors, such as NLR, PLR, mGPS, PNI, and CONUT, [16][17][18][19][20] were not identified as independent prognostic factors. Skeletal muscle changes may be particularly useful, rather than the inflammation-based prognostic parameters mentioned above, in patients with PHC who may develop cholangitis during a long PWT.
As in the previous reports, 6  Recently, besides sarcopenia, there have been reports that sarcopenic obesity is an important prognostic factor for malignancy. 27 However, when limited to PHC, while there are some reports of sarcopenic obesity being associated with postoperative complications, there are still few reports about long-term outcomes. 28 In the present study, sarcopenic obesity was also not associated with OS and RFS in patients with resected PHC due to univariate Cox regression analysis. Further research is needed on the prognostic impact of sarcopenic obesity in patients with PHC.
The mechanisms by which sarcopenia or skeletal muscle wasting is associated with morbidity and mortality are not fully understood.
However, this association could be because more malignant tumors can lead to energy reduction by eliciting an excessive inflammatory response, resulting in loss of muscle mass. 29 In this study, we observed significantly more poorly differentiated cancers in the preoperative skeletal muscle wasting group than in the skeletal muscle no-change group. If the sarcopenia is related to tumor malignancy, methods such as the introduction of intense neoadjuvant therapy However, due to the retrospective design of the study, the evaluation of muscle function, such as handgrip strength, proved challenging. Thus, further studies are warranted to accurately assess the association between preoperative changes in skeletal muscle mass and the prognosis of patients with PHC.
In conclusion, we demonstrated that preoperative skeletal muscle wasting in patients with PHC had a negative effect on survival outcomes.

ACK N OWLED G M ENT
We thank Editage (www.edita ge.com) for English language editing.