Identification of Risk Factors in Postoperative Urinary Dysfunction for Rectal Gastrointestinal Stromal Tumors.

Jun Kataoka: conceptualization, data curation, formal analysis, investigation, methodology, project administration, software, validation, visualization, writing – original draft, writing – review and editing. Yuichiro Tsukada: conceptualization, project administration, supervision, validation, visualization, writing – review and editing. Masashi Wakabayashi: conceptualization, formal analysis, investigation, methodology, software, validation, writing – review and editing. Daichi Kitaguchi: formal analysis, resources, writing – review and editing. Hiro Hasegawa: resources, writing – review and editing. Koji Ikeda: resources, writing – review and editing. Yuji Nishizawa: resources, validation, writing – review and editing. Masaaki Ito: project administration, supervision, writing – review and editing.

All procedures performed on human participants were in accordance with the ethical standards of the institutional and national research committees and the Declaration of Helsinki (1964) and its later amendments or comparable ethical standards. This study was approved by the Institutional Ethics Review Board of the National Cancer Center Hospital East, Japan (Approval no. 2018–100).

We retrospectively reviewed patients who underwent surgery for primary rGISTs with curative intent between January 2008 and March 2022 at the National Cancer Center Hospital East. The analyses included patient characteristics, clinicopathological features of rGISTs, surgical outcomes, prognoses, and postoperative urinary function. This study was approved by the Institutional Ethics Review Board of the National Cancer Center Hospital East, Japan (approval no. 2018–100). All primary rGISTs were preoperatively evaluated by digital rectal examination, colonoscopy, contrast‐enhanced computed tomography (CT), and magnetic resonance imaging (MRI) [ 15 ]. Tumor size was determined as the largest tumor size on MRI. The shrinkage rate was calculated based on the largest tumor size at initial visit and immediately before surgery. The disease status was assessed by summing the tumor size per the Response Evaluation Criteria in Solid Tumors for preoperative targeted therapy [ 16 , 17 ]. Tumor risk grade was assessed following the National Institutes of Health consensus criteria for assigning risk to GISTs [ 18 , 19 , 20 ]. Risk of aggressive behavior was determined using the Miettinen and modified Fletcher criteria [ 21 , 22 , 23 ]. The specific tumor size guidelines for the introduction of preoperative targeted therapy in rGISTs have not been clearly defined in our hospital; however, this decision was made through discussions on a case‐by‐case basis at multidisciplinary team conferences involving oncologists, surgeons, and radiologists. Preoperative targeted therapy was administered for a planned duration of 6 months due to the requirement for maximal response. Meanwhile, for patients with high‐risk factors for rGISTs, the data suggested that at least 3 years of adjuvant treatment was associated with improved relapse‐free survival (RFS) [ 24 ]; therefore, these patients received chemotherapy for at least 3 years postoperatively. The standard treatment for primary rGISTs with no distant metastasis or dissemination was complete surgical excision of the tumor. Organ‐preserving surgery was performed whenever possible. To ensure R0 resection (macroscopically complete resection with negative microscopic margins) and organ preservation, we introduced two‐team laparoscopic surgery combined with trans‐anal total mesorectal excision [ 17 ]. Laparoscopic surgeries for rGIST were introduced to our division in 2008, with the addition of a transanal approach, as necessary. Since January 2017, two‐team surgery that was simultaneously initiated using an abdominal and transanal approach has been performed, thereby necessitating all surgeries in rGISTs to be performed by two‐team surgery at any time [ 17 ]. In the early stages of the introduction of laparoscopic surgery, cases involving tumor size approximately 10 cm post‐ preoperative targeted therapy or where close proximity to pelvic organs (potentially requiring composite resection) were initially managed with laparotomy. However, in recent years, laparoscopic surgery has become the preferred approach for complicated cases. We categorized the degree of autonomic nerve preservation into complete preservation and partial resection. The urinary catheter was removed 5 days postoperatively because some reports recommended so, and urinary function was evaluated at that time [ 14 , 25 , 26 ]. Residual urine volume (RUV) was measured three times after removal of the catheter using ultrasound, and voiding efficiency (VE), which is the ratio of self‐voiding volume to the total voiding volume, was calculated [ 14 , 27 , 28 ]. VE was calculated using the following formula: VE = [total self‐voiding volume/ (total self‐voiding volume + total RUV)] × 100 (%) [ 28 ]. If VE was < 50%, medications were prescribed to the patients to improve voiding, and if VE did not improve to ≥ 50%, intermittent self‐catheterization or reinsertion of urethral balloon was performed. Patients who required intermittent self‐catheterization or urethral balloon catheter at the time of discharge were classified as having postoperative UD. Furthermore, improvement in postoperative UD was defined as the removal of the urethral balloon catheter or discontinuation of intermittent self‐catheterization. Postoperative follow‐up was conducted using standardized medical records, which included the following procedures: physical examination, chest and abdominal CT every 6 months, and colonoscopy at 1, 3, and 5 years postoperatively. Local or distant recurrences were confirmed based on histological findings and imaging studies, such as CT or MRI. Patient characteristics were summarized using descriptive statistics such as median (range) and proportion. To test for significant differences between groups, univariable analyses were conducted using the Mann–Whitney U test for continuous variables and Fisher's exact test for categorical variables. Multivariable analysis was performed using logistic regression analysis with the backward elimination method to investigate the preoperative independent risk factors for postoperative UD. We used only preoperative characteristics because this analysis was performed to investigate preoperative independent risk factors for postoperative UD. We used age (≥ 65 years vs. < 65 years), sex, body mass index (BMI ≥ 25.0 kg/m 2 vs. < 25.0 kg/m 2 ), presence of preoperative targeted therapy, and preoperative tumor size (≥ 50.0 mm vs. < 50.0 mm) as variables to evaluate the association with the postoperative UD. We established a cut‐off value for preoperative tumor size at a median of 50.0 mm. A p ‐value < 0.05 was considered statistically significant. OS was defined as the duration from treatment initiation to the last follow‐up or mortality while RFS was defined as the duration from surgery until either clinical evidence of recurrent or metastatic disease, or the end of follow‐up. OS and RFS were calculated using the Kaplan–Meier method. All statistical analyses were performed using EZR software for R (Saitama Medical Center, Jichi Medical University, Saitama, Japan) [ 29 ].

Overall, 45 patients were included in this study. Patient characteristics are shown in Table  1 . The median preoperative major tumor size and shrinkage rate were 50.0 mm (range, 10.0–110.0 mm) and 32.9% (range, 0%–54.5%), respectively. The tumor size immediately before surgery with preoperative targeted therapy introduction was significantly larger than that before upfront surgeries ( p = 0.02 ) (Figure  1 ). The maximum diameter can be measured by MRI; thus some of them are in the sagittal direction. Four cases fell into this category. We measured the maximum tumor size and corresponding pelvic diameters in both the horizontal and sagittal directions using MRI at the tumor level. (Table  1 ). No cases had undergone pelvic irradiation and a total of seven cases (uterine fibroids three cases, cesarean two cases, endometriosis one case and) were included in our study. All seven cases included females. Patients characteristics. Median [range]. Box plot of preoperative targeted therapy and upfront surgery. The preoperative tumor size is larger in the preoperative targeted therapy group ( n  = 27) than in the upfront surgery group ( n  = 18). The median preoperative tumor size is 37.5 mm (range, 10.0–70.0 mm) in the upfront surgery group and 59.0 mm (range, 25.0–110.0 mm) in the preoperative targeted therapy group, and a significant difference is observed between the groups ( p  = 0.02). Surgical outcomes are shown in Table  2 . Three patients underwent no sphincter‐preserving procedures. One patient opted against sphincter preservation despite a small preoperative tumor size and therefore underwent a laparoscopic Hartmann procedure. Another patient underwent conversion from laparoscopic low anterior resection to open Hartmann procedure due to a large tumor size (70 mm) and significant intraoperative blood loss (5966 mL). The third patient underwent laparoscopic abdominoperineal resection (APR) due to large tumor size (97 mm), even after preoperative targeted therapy. The degree of autonomic nerve preservation was classified as either complete preservation or partial resection. Preoperative tumor size was significantly larger in patients who underwent autonomic nerve partial resection (median, 60.0 mm [range, 34.0–110.0 mm]) than in those who underwent complete preservation (median, 38.0 mm [range 10.0–105.0 mm]) ( p  = 0.004) (Figure  2 ). Figure  3 clearly demonstrates the association between the preoperative tumor size and reduced autonomic nerve preservation rate. The rate of complete autonomic nerve preservation and partial resection in patients with a preoperative tumor size of 50.0 mm as the threshold revealed that the complete preservation rate of patients with tumor size ≥ 50.0 mm preoperatively (37.5%) was lower than that of those with a tumor size < 50.0 mm (76.2%). No significant differences were observed in the postoperative UD with or without trans‐anal approach (taTME), open surgery or laparoscopic surgery, and surgical procedure (LAR, ISR, APR, and Hartmann) in the univariable analysis ( p  = 0.483, 0.143, and 0.192). Surgical outcomes. Median [range]. Box plot of autonomic nerve complete preservation and partial resection. Autonomic pelvic nerve is preserved in 25 (53.3%) patients, whereas it is partially preserved in 20 (46.7%) patients. Comparison of the degree of autonomic nerve preservation and preoperative tumor size shows that preoperative tumor size is significantly larger in patients who undergo autonomic nerve partial resection (median, 60.0 mm [range, 34.0–110.0 mm]) than in those who undergo complete preservation (median, 38.0 mm [range, 10.0–105.0 mm]) ( p =  0.004). Stacked bar chart of relationship among preoperative tumor size, autonomic nerve preservation rate and postoperative urinary dysfunction in surgery for rectal gastrointestinal stromal tumor. Complete preservation rate of patients with tumor size ≥ 50.0 mm preoperatively (37.5%) is lower than that of those with a tumor size < 50.0 mm (76.2%), furthermore 22.2% of patients with tumor size ≥ 50.0 mm preoperatively suffered postoperative urinary dysfunction even if autonomic nerve was completely preserved. The rate of severe postoperative complications (Clavien–Dindo criteria III or IV) was 21.4% (Table  3 ). Postoperative outcomes, pathological findings, and urinary findings. Median [range]. Pathological findings, including Miettinen and the modified Fletcher classifications, are shown in Table  3 . Two patients could not be classified due to a complete pathological response to treatment (4.4%). The 5‐year OS and RFS rates are shown in Figure  S1A,B . Of the 45 patients, 44 underwent R0 resection. In R1 resection, we initially performed laparoscopic intersphincteric resection with trans‐anal approach (taTME) and prostate composite resection by these patients with simultaneous liver metastases (44‐year‐old male patient, imatinib therapy for 8 months preoperatively). Tumor was located on the anterior side and at 3.0 cm from anal verge. The operation time was 429 min, including the liver resection time (our part: 282 min), and blood loss was 80 mL. We performed curative resection based on the intraoperative findings; however, it was later discovered pathologically to have an R1 resection. As a second surgery, we performed total pelvic exenteration to ensure complete resection of the remnant after approximately 2 months from the initial surgery. Fortunately, no recurrence was observed after an additional resection at 31 months after second surgery. Among the three cases of recurrence, one was local recurrence at 45 months postoperatively requiring reoperation (transabdominal surgery); this patient had no recurrence thereafter. Two cases of recurrence included distant metastases to the bone (diagnosed at 43 and 37 months postoperatively). The patient with recurrence at 43 months died at 82 months due to disease progression. Eleven patients (24.4%) had postoperative UD at discharge. The median duration for postoperative UD recovery was 141 days (range, 43–471 days). Among 11 patients with postoperative UD, seven received pharmacological treatment (parasympathomimetic agents and α‐blockers). Lower urinary tract is dominated by three types of peripheral nerves: pelvic, hypogastric and pudendal nerves; moreover, the pelvic and the hypogastric nerves contribute to the development of UD following pelvic surgery. In our analysis, 20 cases of partial autonomic nerve preservation were evaluated (14 cases of unilateral total preservation and six cases of bilateral partial resection), including one case in which the unilateral hypogastric nerve was also transected. Table  4 summarizes the results of univariable and multivariable analyses of postoperative UD and preoperative characteristics. In multivariable analysis, postoperative UD was significantly associated with tumor size ≥ 50.0 mm immediately before surgery (odds ration [OR] = 8.96, 95% confidence interval [CI]: 1.46–55.20; p  = 0.02) and male sex (OR = 8.16, 95% CI: 1.32–50.60; p  = 0.02). Figure  4 shows the incidence of postoperative UD by surgical procedure. There were a small number of no sphincter‐preserving procedure cases (one patient who underwent APR and two patients who underwent Hartmann procedure), however there was no statistically significant difference under subgroup analyses by surgical procedure due to the small sample size. Meanwhile, there was no association with the occurrence of postoperative UD and severe postoperative complication in the univariate analysis. ( p  = 0.705). Similar, no association in the occurrence of postoperative UD and pelvic surgery in the past was observed in the univariable analysis ( p  = 0.663). Table  S1 summarizes identification of independent risk factors for postoperative UD while based on the preoperative and intraoperative factors against Table  4 . Preoperative risk factors of urinary dysfunction at the time of discharge according to univariable or multivariable analyses. Fisher's exact test. Mann–Whitney U ‐test. Multivariable logistic with backward elimination method. Stacked bar chart of subgroup analyses by surgical procedure for rGISTs. The breakdown of rGIST procedures performed at our hospital is LAR, 23; ISR, 19; APR, 1; and Hartmann procedure, 2. There is no significant difference in the development of postoperative urinary dysfunction between different surgical procedures ( p  = 0.192).

In this study, we investigated the independent risk factors for postoperative UD during surgery for rGISTs. Among the 45 patients with rGISTs, 44 underwent curative surgery, with 5‐year OS and RFS rates of 100% and 88.5%, respectively, confirming that the outcomes in our study were comparable to numerous studies in terms of the highly curative outcomes [ 10 , 30 , 31 ]. In terms of postoperative quality of life, the incidence of postoperative UD in patients with low rectal cancer was 18.5% in our division [ 14 ], whereas in this study, among 45 patients, 11 had postoperative UD (24.4%). One case with R1 resection was initially performed using laparoscopic low anterior resection with taTME, who experienced postoperative UD at discharge. Then, we performed total pelvic exenteration as secondary surgery owing to pathological finding suggesting R1 resection. Therefore, postoperative UD occurred after the initial surgery and before second surgery, this patient was included in our analyses, which based on the initial outcome; thus, we considered this approach appropriate for postoperative UD. Multivariable analysis for identifying preoperative risk factors for postoperative UD revealed that male sex and tumor size ≥ 50.0 mm immediately before surgery were independent risk factors for postoperative UD. Only the multivariable analysis revealed that male sex was a preoperative independent risk factor for postoperative UD as there might be a slightly significant difference in male sex between the postoperative UD and non‐UD groups ( p  = 0.08 in univariable analysis). Since male sex was not a clear confounding factor, it could be considered a preoperative independent risk factor for postoperative UD. In surgery for rGISTs, achieving complete removal of the tumor‐bearing rectal wall and tumor‐covering tissue layer is critical, as rGISTs originate from the muscularis propria. rGISTs do not infiltrate adjacent pelvic organs; therefore, local resection along the tumor is preferred for the functional preservation of the organ, as preservation of pelvic organs and nerves is ideally possible. However, it is difficult to secure clear resection margins when rGISTs are large, occupied in pelvic space and in close proximity to various pelvic adjacent organs, such as the prostate, levator ani muscle, and vagina. Furthermore, it was easy to intraoperatively damage the anal sphincter and peripheral autonomic nerve, posing a high risk of infection, fistula, short‐term bleeding, and possible postoperative sexual, urinary, and defecation dysfunction in the long term [ 32 , 33 ]. To avoid these, extended radical operation such as APR and pelvic exenteration (PE) had been performed [ 11 ]. Therefore, preoperative targeted therapy introduction has recently been considered for reducing the tumor size [ 34 ]. The median tumor shrinkage rate with preoperative targeted therapy is reported to be 33% [ 35 ], leading to an increased rate of urinary tract and sphincter preservation [ 10 , 11 ]. Some studies showed that although APR was initially planned for 47.8% of patients, it was performed in only 30.4% of patients following preoperative targeted therapy; similarly, total PE was initially planned for 34.8%, but was performed in 21.7% of patients due to preoperative targeted therapy effects [ 11 ]. Our study showed that the shrinkage rate in patients who underwent preoperative targeted therapy was 32.9%, and preoperative targeted therapy led to urinary tract preservation in some patients. Our previous analysis was the only study to explore postoperative UD in patients with rGIST [ 14 ]. Postoperative main outcomes of urinary function in rectal cancer are predominantly attributed to the damage of the pelvic autonomic nerve [ 15 , 16 ]. However, surgery for rGIST is based on resection along the tumor to avoid damage to the autonomic nerve and iliac vessels as much as possible, and if the autonomic nerves are resected, it is expected that the tumor size is large and severely close to them. Furthermore, our study showed that the preoperative tumor size with autonomic nerve and complete preservation was significantly larger than that with partial nerve resection ( p =  0.004), and the preoperative tumor size above the threshold (50.0 mm) had a lower rate of complete preservation than those below. Therefore, as shown in Figure  3 , tumor size ≥ 50 mm was associated with a markedly reduced rate of complete nerve preservation, suggesting that nerve preservation becomes technically challenging when the tumor occupies significant pelvic space. Traditionally, rGISTs have been considered to completely preserve the nerve due to limited invasion of the surrounding organs and structures. However, our study showed that as the tumor size increased, the pelvic space for surgery gradually narrowed, and the tumor approached the surrounding nerve, making complete preservation of the autonomic nerve complicated and increasing the risk of postoperative UD. Incidentally, in our analyses, no significant differences were observed in the postoperative UD incidence among different surgical procedures (LAR, ISR, APR, Hartmann) and with or without taTME. Our findings revealed the trend toward greater difficulty with nerve preservation in patients with tumor size ≥ 50.0 mm immediately before surgery, suggesting that tumor reduction with preoperative targeted therapy may better preserve urinary function, especially in patients with initial tumor size ≥ 50.0 mm (before preoperative targeted therapy). However, this study had certain limitations. It was a retrospective analysis performed at a single center, which may limit the generalizability of the findings. In addition, the confidence intervals are wide due to the small number of postoperative UD onset events. However, in our study, statistically significant differences and large ORs were obtained even under this situation; therefore, we considered that male sex and tumor size ≥ 50.0 mm immediately before surgery were relevant risk factors to be aware of in postoperative UD in a series of clinical perspective. Compared with our previous report [ 17 ], this study included a larger sample size and provided a comprehensive analysis, including quality of life evaluation, pathological findings, and risk factors for postoperative UD. None of the studies, including our previous study, evaluated postoperative quality of life. Furthermore, our study concluded that tumor size ≥ 50.0 mm immediately before the study was a preoperative risk factor for postoperative UD may aid in the introduction of preoperative targeted therapy. Our study suggested that tumor reduction through preoperative targeted therapy may contribute to improved postoperative quality of life. This approach differs from conventional rectal cancer surgeries, emphasizing the unique aspects of rGIST treatment.

The incidence of UD after surgery for rGISTs was 24.4%. The preoperative risk factors for UD were male sex and tumor size ≥ 50.0 mm immediately before surgery. Tumor reduction through preoperative targeted therapy may better preserve urinary function, particularly in cases with initially large tumors (≥ 50.0 mm).

Gastrointestinal stromal tumors (GISTs) are the most common types of mesenchymal soft tissue tumor, originating from the interstitial cells of Cajal in the gastrointestinal tract, with an incidence of 10–15 per million per year [ 1 , 2 ]. Most (75%–80%) GISTs possess oncogenic gain‐of‐function mutations in the c‐Kit ( CD117 ) gene, while approximately 10% express platelet‐derived growth factor receptor alpha mutations [ 3 , 4 , 5 ]. GISTs present throughout the gastrointestinal tract, with the highest occurrences in the stomach (60%–70%) and small intestine (20%–30%), following the colon and rectum (5%–10%) [ 6 ]. Curative resection is the first‐line treatment for primary GISTs, with a 5‐year overall survival (OS) rate of 50%–80% [ 7 , 8 ]. Among 20 studies and reviews on rectal GISTs (rGISTs), R0 (macroscopically complete resection with negative microscopic margins), R1 (microscopically positive margins), and R2 (macroscopically incomplete resection) resections were achieved in 80/107 (74.7%), 26/107 (24.2%), and 1/107 (9.35%) patients, respectively; postoperative recurrence occurred in 3/80 (R0: 3.75%), 6/26 (R1: 23.1%), and 1/1 (R2: 100%) patients [ 9 ]. Furthermore, severe postoperative complications developed in 21.4% of cases [ 10 ]. rGISTs lie in close proximity to adjacent pelvic organs, and the preservation of organs, such as the bladder, prostate, uterus, and vagina, is challenging. Therefore, we introduced preoperative treatment with imatinib to reduce the size of the tumor and achieve radical cure and organ preservation. As a result, previous studies on preoperative targeted therapy with pre‐imatinib era reported urinary tract and sphincter‐preserving surgery rates of 28.5% and 54.8%; however, some other reports on preoperative targeted therapy with imatinib demonstrated an increase in the rate of urinary tract and sphincter‐preserving surgery (33.3%–100%) [ 11 , 12 , 13 ]. Organ preservation, including urinary tract, is a key consideration in postoperative quality of life. However, despite the increased rate of preservation, the only reports on postoperative quality of life are for postoperative rectal cancer [ 14 , 15 , 16 ]. Our previous studies on postoperative urinary function demonstrated that the frequency of postoperative urinary dysfunction (UD) in rGISTs was 23.8% [ 17 ]. As opposed to rectal cancer, rGIST surgery generally does not require lateral lymph node dissection or preoperative radiotherapy; furthermore, tumors can often be dissected along the capsule, enabling preservation of adjacent urinary structures and, theoretically, autonomic nerves. However, when the tumor is large, even such an approach may not permit complete autonomic nerve preservation, leading to UD. These differences underscore the need for investigating risk factors for postoperative UD in rGISTs, distinct from rectal cancer; therefore, this study aimed to retrospectively assess the risk factors for postoperative UD in rGISTs.

The authors declare no conflicts of interest.

FIGURE S1. Kaplan–Meier curves for overall survival (OS) and relapse‐free survival (RFS) of patients with rectal gastrointestinal stromal tumor. FIGURE S2. Box plot of preoperative tumor size ‐to‐pelvic space ratios in sagittal direction between patients with and without postoperative urinary dysfunction. FIGURE S3. Box plot of preoperative tumor size‐to‐pelvic space ratios in horizontal direction between patients with and without postoperative urinary dysfunction. FIGURE S4. Box plot of pelvic size in sagittal direction between sexes. FIGURE S5. Box plot of pelvic size in horizontal direction between sexes. TABLE S1. Independent risk factors for postoperative UD while based on the preoperative and intraoperative factors.