A total of 68 patients with sizeable retroperitoneal tumors, defined as tumors with boundaries beyond the midline of the abdomen (over the midline of the spine on imaging), were enrolled in the present study (Table I). A curved incision along the costal margin, or in combination with a longitudinal incision along the tumor boundary and the lower boundary of the tumor along the midline of the abdomen, were used for patients in group A (32 patients, male:female ratio 17:15, including 9 children), with a mean age of 56.7±12.8 years (range, 22 months-74 years). The longest diameter of the tumors ranged from 14 to 43 cm, with a mean ± standard deviation (SD) of 21.3±4.4 cm. Of the 32 patients, 17 had tumors on the left side and 15 on the right side. An abdominal incision or an oblique incision at the waist were used in group B, which included 36 patients (male:female ratio 19:17, including 11 children) with a mean age ± SD of 54.2±12.1 years (range, 20 months-74 years). The longest diameter of the tumors ranged from 12 to 43 cm, with a mean ± SD of 22.3±3.4 cm. Of these patients, 17 had tumors on the left side and 19 on the right side. A total of 14 patients (20.6%) had benign and 52 (69.4%) had malignant tumors. The clinical manifestations included abdominal pain in 22 cases, bloating in 34 cases, nausea, frequent or urgent urination and ipsilateral lumbar bulging in certain cases, whereas 12 patients were asymptomatic. The tumors in all the patients were located in the abdomen and across the midline of the abdomen. Preoperative computed tomography (CT)/magnetic resonance imaging (MRI) were used to measure the tumors and revealed that the tumors extended beyond the midline of the spine, exerting pressure on the gastrointestinal tract in 64 cases. Isotope renography was used to assess renal function in 60 cases, with 38 cases displaying renal and ureteral compression or obstruction and 6 cases of renal dysfunction. The function of the contralateral kidneys was normal. CT and MRI revealed abdominal aortic pressure and displacement in 21 cases, invasion and encapsulation in 11 cases, abdominal venous pressure and displacement in 32 cases, infiltration and encapsulation in 13 cases and encapsulation of the renal artery and vein in 15 cases. There were no significant differences between patients in groups A and B in terms of gender, tumor size, extent of infiltration around the left and right side, age and postoperative pathology. As regards preoperative preparation, 1,100 ml blood was prepared on average per surgery and relevant consultations and coordination between departments were required to complete surgery in 15 cases.

For patients in group A, surgery was performed via a costal margin curved incision; a longitudinal incision was made along the tumor boundary and down to its lower boundary at the midline of the abdomen, as well as an incision at the rib outside the ipsilateral axillary line. When possible, the curve angle was not <90 degrees. For cases of superior renal and adrenal tumors, the prone position was used and the patients were elevated by 15 degrees (15 cases in total). For all other patients, the supine position was used. Following incision of the skin and abdominal fat layer, the muscle was separated at different levels to facilitate the identification of layers when suturing the incision, thus reducing the incidence of postoperative incisional hernia. The operative field was fully visualized and the anatomy and organizational associations were identified (1–10). In a region with relatively few blood vessels, a lateral peritoneal incision was performed outside the colon and the tumors were carefully separated. For retroperitoneal tumors, the surgical approaches may differ depending on the side of the tumor; the degree of complexity is higher in retroperitoneal tumors located on the right side, where important and easily-damaged organs, such as the liver, duodenum and vena cava, are located. In addition, the surgical field is more restricted. As a result, the most appropriate treatment was performed in conjunction with abdominal surgery with collaboration across multiple disciplines. For left-sided tumors, the pancreas, spleen with associated blood vessels and the aorta were identified and located prior to surgery. For tumors close to the mesenteric root, every effort was made to prevent damaging the superior mesenteric artery and vein.

For patients in group B, the incision was made at the midline of the abdomen or obliquely along the waist. The requirement for an additional incision for the purpose of obtaining the optimal surgical field was determined following abdominal exploration. If the tumor was too large and difficult to directly observe, the tumor and surrounding organs were assessed via the abnormal anatomical associations. The methods and techniques for isolating the tumors were the same as those used in group A patients.

A statistical analysis was performed using SPSS software, version 20.0 (IBM Corp., Somers, NY, USA) and the analysis of continuous variables between the two groups was performed using the Mann-Whitney U test. The comparison of categorical variables between the two groups was performed using the Chi-square test and the significance level was set at 0.05.

Surgical resection was performed in the 32 patients in group A, of whom 22 underwent complete tumor resection, 1 underwent resection of the tumor capsule, 6 patients underwent sub-block resection, whereas complete resection with excision of adjacent organs was performed in 3 cases. In the cases undergoing complete resection, 1 patient required repair following injury of the vena cava and right renal vein and 1 patient required vena cava and aortic damage repair. Among the 36 cases in group B, surgical resection was performed in 25 patients, with complete resection achieved in 19 patients; resection of the tumor capsule was performed in 1 patient, sub-block resection was performed in 3 patients and removal of the tumor and surrounding organs was required in 2 patients. Surgical resection failed in 11 cases (6 patients were only subjected to pathological examination, 3 patients had >50% residual tumor and 2 patients had <50% residual tumor).

The data regarding transfusion and operative time are presented in Table II. In group A, blood transfusion was performed in 19 cases, with a mean ± SD of 400±200 ml blood (range, 600–1,900 ml). The longest operative duration was 4.5 h and the shortest 2 h, with a mean ± SD of 3.5±0.5 h. The weight of the resected tumor ranged between 3.5 and 10 kg, with a mean ± SD of 4.3±0.8 kg. In group B, surgical tumor removal was performed in 30 patients (6 patients only received pathological examination) and 25 patients were transfused with a mean ± SD of 1,100±200 ml blood (range, 400–4,500 ml). The mean ± SD operative duration was 5.8±0.7 h (range, 10–3.5 h) and the weight of the resected tumor varied from 1.2 to 7.9 kg, with a mean ± SD weight of 2.4±0.7 kg.

There were no reported deaths or development of intestinal fistulas in group A. The patients were discharged from the hospital after a mean of 8.5 days after surgery. In group B, 1 patient succumbed to pneumonia and intestinal fistulas developed in 2 patients. The patients were discharged from the hospital after a mean of 11.5 days after surgery (Table II).

Variables such as patient age, location of the tumor and tumor size were not statistically significantly different between groups A and B (P=0.274, 0.138 and 0.224, respectively; Table II), indicating that the basic characteristics of the patients were comparable between the two groups. However, there were statistically significant differences regarding operative time, volume of transfused blood, postoperative in-hospital stay, weight of surgically removed tumor, perioperative mortality, development of intestinal fistula after surgery and surgical resection rate in any pairwise comparison (Table II), indicating that a curved incision under the costal margin was superior to the abdominal midline or oblique waist incision, in terms of surgical resection rate and safety.