Complications and Outcomes of Diaphragm Surgeries in Epithelial Ovarian Malignancies

Abstract

To achieve optimal debulking, cytoreductive surgery often involves diaphragm stripping. We describe our complications and survival outcomes after diaphragm surgery in epithelial ovarian cancer. A retrospective analysis on patients with advanced stage epithelial ovarian cancer between January 2012 and September 2019. The details of the diaphragmatic resections and stripping and their complications were looked into. During the study period, 616 patients with epithelial ovarian cancers were operated of which, 81 (13.2%) had diaphragm surgery. The majority underwent diaphragm stripping (60%) while 33% had resection and 7% cases had diaphragmatic nodule excision. Optimal debulking was achieved in 89% of cases. The complexity of surgery was intermediate in 64% of patients and complex in 33% as per Aletti’s scoring. Mean operating time was 300 min (SD113). Moderate to severe pleural effusion was seen in 26 (32. %) patients necessitating pleural tapping in 16% and single lumen pleurex catheter insertion in 11%. Median recurrence-free and overall survival were 22 (95% CI 16.9–27) and 32 months (95% CI 25.5–38) respectively. Diaphragm stripping and resection is an important step in achieving optimal debulking of advanced and recurrent ovarian cancer. Diaphragmatic disease clearance is a necessary skill to be acquired by the gynaecologic oncology surgeons. Choosing the patients correctly and anticipation of complications can reduce morbidity and mortality

Introduction

Ovarian cancer has been on a rise all over the world with the GLOBOCAN data estimating the age-adjusted incidence and mortality rates to be 6.6 and 3.9 per 100,000 women across the globe [1]. Almost 70% are detected in advanced stages, quite often involving the upper abdomen and diaphragm [2, 3]. Primary cytoreductive surgery followed by adjuvant chemotherapy or neoadjuvant chemotherapy with platinum or taxane compounds is considered as the standard of care for advanced ovarian cancer [4, 5].

Cytoreduction is considered to be complete when there is no macroscopic residual disease, optimal when there is disease up to 1 cm residual disease and suboptimal if gross residual disease is > 1 cm. Complete resection of all macroscopic disease has been shown to be the single most important independent prognostic factor affecting outcome in advanced ovarian cancer [4].

About 40% of patients with advanced ovarian cancer have diaphragmatic involvement and require liver mobilization with diaphragm peritonectomy or resection to achieve complete cytoreduction. Splenectomy and distal pancreatectomy may be required in some cases to achieve complete cytoreduction of diaphragmatic disease. Superficial extension of disease requires stripping of the peritoneum, whereas diaphragmatic resection is performed if there is a full thickness or partial muscle infiltration [6–8]. Complete cytoreduction by stripping and resection of the diaphragm seem to improve the overall survival of patients [9] but may be associated with an increase in intra-operative as well as post-operative morbidity [10].

This study was conducted primarily to evaluate diaphragmatic surgery in the setting of cytoreduction for advanced epithelial ovarian malignancies. We also wanted to assess the peri-operative as well as post-operative complications of the patients undergoing diaphragmatic procedures as part of the cytoreductive surgeries.

Materials and Methods

The study was conducted in a gynaecologic oncology department at a tertiary care centre in India. It was a retrospective analysis of all patients who had undergone diaphragm surgery as part of the debulking procedure for advanced ovarian cancer. Approval was obtained from the institutional review board.

Participants

We reviewed the medical records of all patients who underwent surgery for advanced epithelial ovarian cancers (International Federation of Gynaecology and Obstetrics [FIGO] stages IIIC/IV), from January 2012 to September 2019. All patients with advanced epithelial ovarian malignancies undergoing cytoreductive surgeries with diaphragmatic surgery as part of their primary, interval, completion and secondary cytoreduction were included in the study. Patients with non-epithelial ovarian malignancy or with synchronous malignancies were excluded from the study. Patients who underwent cytoreductive surgeries without diaphragmatic procedure were also excluded.

Data Collection

Clinical details were retrieved from online electronic medical records. Demographic details, pre-operative investigations, FIGO staging at the initial visit, histopathology of the tumour, intra-operative details, surgico-pathological characteristics, adjuvant therapy, and follow-up details were extracted. Data regarding the complications specific to diaphragmatic resection and general complications of the debulking surgeries were also retrieved.

Departmental Protocol

Pre-operative workup included complete blood counts, renal, and liver function tests. Baseline serum CA 125 as well as other markers like CEA and CA 19–9 was performed where indicated. All the patients had a plain radiograph of the chest, electrocardiograph, and a contrast-enhanced computed tomography of the abdomen and pelvis. Those who presented with recurrence had a PET-CT. A multidisciplinary tumour board evaluated the feasibility of surgical resection and decided on further management based on these investigations.The treatment of choice was primary cytoreduction. However, neoadjuvant chemotherapy was offered in patients with poor performance status (ECOG 3 or more), and R0 resection was not feasible due to encasement of major vessels, involvement of root of mesentery, liver or lung parenchyma. Additionally in our setting, neoadjuvant chemotherapy was also considered if the patient refused extensive surgery with bowel resection and stoma or if surgery could not be scheduled within 3 weeks of presentation due to extensive case load.

In such cases, patients received 3 to 4 cycles of paclitaxel and platinum-based chemotherapy prior to surgery every 21 days. Interval debulking was offered after 3 cycles of chemotherapy. An attempt to achieve an R0 resection was made both at upfront and interval surgery. Removal of diseased diaphragm was mandatory to achieve this. Completion surgery was done in those patients who had incomplete surgeries outside, mainly after ovarian cystectomy or total abdominal hysterectomy with bilateral salpingooophorectomy. Secondary cytoreduction was offered to patients of good performance status who had platinum sensitive recurrences and had already received 2nd- or 3rd-line chemotherapy. The surgery was performed only if the disease was localised and complete resection was likely. Hyperthermic intra-peritoneal chemotherapy (HIPEC) was given in few selected cases of interval and secondary cytoreduction.

The procedure began with exposure of the midline incision upto the xiphoid process. Omnitract retractors were used to lift the costal margin. The falciform ligament was dissected initially, followed by the right coronary and triangular ligaments. This ensured complete hepatic mobilization by separating the liver from its peritoneal attachments with the diaphragm and the posterior abdominal wall to allow adequate exposure of the right hemidiaphragm. After thorough exploration, the procedure of diaphragmatic cytoreduction was decided upon depending on the diaphragmatic involvement by the tumour. Intra-operative conditions like haemodynamic instability of the patient secondary to liver mobilization were also taken into account while selecting the procedure. Small lesions mainly in the form of nodular/localised involvement of the diaphragm were resected with the help of electrocautery. Stripping, that is dissection of the peritoneum of the affected area, was considered in cases of disseminated but superficial diaphragmatic peritoneal involvement by tumour. Extensive disease with full-thickness involvement penetrating the underlying muscle and pleura were managed with diaphragmatic resection. Monopolar cautery was used during the procedure. To perform diaphragmatic stripping, the involved peritoneum was separated from the underlying musculature and after holding the edges with multiple sponge holding clamps, the affected area was dissected off. In those areas which necessitated diaphragmatic resection, the lesion was caught with long Allis clamps and the margins were resected sharply. Care was taken to avoid injury to major vascular structures like hepatic veins. After resection, diaphragm was sutured with a non absorbable suture like 1–0 Prolene. A nasogastric tube which was connected to a suction apparatus was inserted into the pleural space before the terminal suture. The catheter was gradually withdrawn to evacuate the pneumothorax while the anaesthesiologist gave the patient maximal inspiration and positive expiratory pressure was maintained. The diaphragm was inspected for leaks by a Valsalva manoeuvre and checked for obvious evidence of significant residual pneumothorax. We closed the diaphragm was closed primarily in all cases and did not use mesh in any patient. Only in cases of pleural tear or intra-op lung collapse ICD insertion was considered during the surgery. All the patients routinely except the ones with diaphragmatic nodule excision and minimal stripping underwent a plain X-ray of the chest on post-operative day 2. In patients being shifted to ward, bilateral air entry and need for oxygen requirement were monitored. In cases of mild hypoxemia, patients were managed with free flow oxygen by nasal prongs or face mask. Patients shifted to ICU were screened using ultrasound and cases developing moderate pleural effusion (more than 500 ml causing hypoxemia that is partial pressure of oxygen < 60 mm Hg) were considered for pleural tapping/single lumen 7 French Catheter insertion by intensivists/pulmonologists as per institutional protocol. ICD insertion was considered mostly in cases of respiratory distress unrelieved by pleural tapping.

Statistical Analysis

For continuous data, descriptive statistics like mean (SD) was calculated. Frequency and percentages were determined for categorical data. Regression analyses were used to determine the association between different variables. All the tests were 2-sided and a p value of < 0.05 was considered significant. The overall survival (OS) of the patients was calculated as the time (in months) from the date of surgery or initiation of chemotherapy to the date of death or last follow-ups. Disease-free survival (DFS) was calculated as the time (in months) from the date of surgery or initiation of chemotherapy to the documented first recurrence. All analyses were done using Statistical Package for Social Services (SPSS) software Version 21.0 (Armonk, NY; IBM Corp).

Results

During the study period, 616 patients with epithelial ovarian cancer were operated on of which, 81 (13.2%) patients had diaphragm surgery. Among the study population, 34 patients (42%) had primary debulking, 36 patients (44%) had interval cytoreduction, while 8 (9.9%) of them had secondary cytoreduction and 3 (3.7%) patients had completion surgeries. Out of 3 patients, 2 had total abdominal hysterectomy with bilateral salpingo oophorectomy (TAH BSO) and omentectomy after 35–45 days of the initial surgery. The other one had received chemotherapy after initial TAH BSO and underwent completion after 3 cycles of chemotherapy. The mean age was 48.16 years (SD 10.7) while the mean BMI was found to be 26.08 kg/m² (SD 4.2). Median serum CA 125 level was 450 U/ml (range 63–6300). The pre-chemo CA 125 was considered in cases receiving neoadjuvant chemotherapy. Serum albumin values had a mean of 3.79 g/dl (SD 0.56). Majority of the patients had stage IIIC tumours and a high-grade serous histology (Table 1).

Table 1.

Baseline characteristics of patients with epithelial ovarian cancer undergoing diaphragm surgery (N = 81)

	N = 81
Age (mean)	48.16 (SD 10.7)
BMI (kg/m2)	26.08 (SD 4.2)
CA 125 mean (U/ml)	450 (range 63–6300)
Stage of disease
IIIc	51 (62.9%)
IVa	16 (19.8%)
IVb	6 (7.5%)
Recurrent disease	8 (9.8%)
Duration of surgery (min)	300 (SD 113)
Blood loss (ml)	1418 (SD 753)
Histology
High-grade serous	70 (86.4)
Low-grade SC	2 (2.5)
Mucinous	6 (7.4)
Endometroid	3 (3.7)

Optimal debulking (less than 1.0 cm residual disease) was achieved in 72 patients (89%) while 9 (11%) had suboptimal debulking. Mean blood loss during surgery was 1.4 l. The mean duration of surgery was 300 min (SD113), range (150–720 min).

Diaphragm involvement was on right side in majority of patients, 57 (70%), while it was bilaterally involved in 22 (27%) of them. Excision of tumour nodules (group 1) was performed in 5.1%, diaphragmatic stripping (group 2) in 60.5%, and resection of the diaphragm (group 3) in 33.4%. None of the patients needed a mesh or diaphragmatic grafting.

Most patients had extensive cytoreduction that involved parietal peritonectomy, rectosigmoid resection and anastomosis, ileal resection and anastomosis, splenectomy, cholecystectomy, and distal pancreatectomy. Majority of the patients, 54/81 (64%) had intermediate complexity of surgery by Aletti’s scoring, while 27 (33%) had highly complex surgery. Splenectomy was performed along with diaphragmatic procedure in 18 (22.2%). Two-thirds (74%) of the patients had peritonectomies and about half (49%) the patients had bowel resection with anastomosis (49%) as shown in Table 2. Figure 1 shows a full thickness resection specimen of the diaphragm.

Table 2.

Details of diaphragm surgeries for epithelial ovarian cancer (N = 81)

Types of surgery	N (%)
Primary debulking surgery	34 (42.0)
Interval debulking surgery	36 (44.4)
Secondary cytoreduction	8 (9.9)
Completion	3 (3.7)
Types of diaphragmatic surgery
Nodule excision	5 (6.5%)
Stripping	49 (60.1%)
Resection	27 (33.4%)
Other surgeries
Peritonectomy	60 (74.0)
Large bowel resection	30 (37.0)
Small bowel resection	9 (11.0)
Stoma	24 (29.0)
Splenectomy	18 (22.2)
Hemicolectomy	3 (3.7)
Cholecystectomy	4 (4.9)
Distal pancreatectomy	4 (4.9)
Partial hepatic resection	2 (2.4)

PDS primary debulking surgery, IDS interval debulking surgery

Fig. 1.

Specimen of diaphragmatic resection

Nearly 70 patients (86%) needed an intra-operative blood transfusion while around 39 patients (48%) required a transfusion in the immediate post-op period. HIPEC was done in 9 (11%) patients.

Thirty-two (39%) patients were admitted to the intensive care unit for monitoring. Eight patients (7%) needed re-admission to ICU in view of worsening sepsis or unresolved respiratory distress. Mean ICU stay for the study cohort was 2.1 days (SD 2.5) while mean hospital stay was 9.6 days (SD 7.5). The commonest post-operative complication was moderate to severe pleural effusion in 27 (35%) patients. In these patients, single lumen ICD insertion was needed in 10 (34.5%) patients in order to drain pleural fluid, improve respiratory efforts, and relieve dyspnoea while pleural tap was sufficient for 15 (55%) subjects while two were managed conservatively. Intra-operatively ICD was placed in for one patient as there was air leak after Valsalva manoeuvre post-closure of defect. Diaphragmatic paresis was suspected in 2 patients and was confirmed in one after fluoroscopy. Both patients had undergone diaphragmatic resection as part of the cytoreduction and were shifted to ICU post-operatively for monitoring. The type of diaphragmatic cytoreduction performed was not found to be significant in predicting the occurrence of post-operative pleural effusion in our study.

Re-laparotomy was required in 4 (4.9%) patients. The other complications included pelvic collection and pigtail catheter insertion in 3 (3.7%), paralytic ileus (12%) wound infection (7.4%), septicaemia (4.9%), and urinary tract infection (4.9%). Mean duration of antibiotics was 5 days in the 29 (35%) patients who required it for probability of sepsis while the rest had only 2 prophylactic doses (Tables 3 and 4). The complication scores according to Clavien-Dindo scoring are graded in Table 4.

Table 3.

Complications (N = 56)

Complications	N (%)
Pulmonary (29/81)
Moderate to severe pleural effusion	20 (24.7)
Moderate to severe effusion with pneumonia	2 (2.5)
Moderate to severe pleural effusion with sepsis and pelvic collection	4 (4.9)
Moderate to severe pleural effusion with sepsis and diaphragmatic paresis	1 (1.2)
Pneumothorax	2 (2.5)
Non-pulmonary (27/81)
Pelvic collection	3 (3.7)
Paralytic ileus	10 (12.0)
Septicaemia	4 (4.9)
Wound infection	6 (7.4)
UTI	4 (4.9)

Table 4.

Classification of complications according to Clavien-Dindo scale (N = 165)

Complications	N (%)
Grade 1
Wound infection (not requiring resuturing)	2 (2.4%)
Pleural effusion/pneumothorax on conservative management	4 (4.9%)
Paralytic ileus (managed by observation/electrolytes/laxatives)	6 (7.4%)
Grade 2
Post-op blood transfusion	39 (48%)
Antibiotics	33 (40.7%)
Total parenteral nutrition	20 (24.6%)
Paralytic ileus (needing IV prokinetics)	4 (4.9%)
Grade 3
3a
Wound resuturing	4 (4.9%)
Pleural effusion requiring drainage	25 (30.8%)
Pigtail insertion for pelvic collection	4 (4.9%)
3b
Relaparotomy	6 (7.4%)
Grade 4
4a
Unresolved respiratory distress requiring intubation	3 (3.7%)
Mechanical ventilation	2 (2.4%)
Tracheostomy	1 (1.2%)
4bnil
Multiple organ dysfunction	6 (7.4%)

In our study, 44 (54.3%) of the patients had developed recurrence. Majority (71.6%) of the patients were alive at the time of reporting. The median overall survival of the study population was 32 months (95% CI 25.5–38) while the recurrence-free survival was found to be 22 months (95% CI 16.9–27) (Table 5).

Table 5.

Interventions done for complications (N = 32)

Interventions	N (%)
Pleural tap	13 (16.0)
ICD insertion	9 (11.0)
Pigtail insertion	3 (3.7)
ICD with relaparotomy	2 (2.5)
Pleural tap with pigtail insertion	1 (1.2)
Relaparotomy	4 (4.9)

Discussion

A key issue in patients with advanced epithelial ovarian cancer is to achieve complete surgical cytoreduction. Resection of all visible disease should be the goal. The earliest reports of optimal debulking and improved in survival in patients with ovarian carcinomas reported better survival if the diameter of residual tumour implants after debulking did not exceed 15 mm [11]. Since then, several studies have emphasised the importance of achieving optimal cytoreduction (residual tumour < 1 cm) to improve survival [5, 12, 13]. The diaphragm is one of the most frequently involved metastatic sites in the upper abdomen in patients with carcinoma ovary. Although 40% of the patients with advanced ovarian cancer present with metastatic diaphragm deposits [8], up to 77% of bulky tumours extending beyond the pelvic brim with involvement of the diaphragm have also been reported [14]. Therefore, achieving optimal cytoreduction would entail diaphragmatic debulking. Diaphragmatic debulking had been deemed as a desirable skill to be acquired during the course of training of a gynaecologic oncology surgeon [15]. Hence, it is of utmost importance that one is well aware of the various techniques involved and the possible complications associated with the procedure.

In a survey done in 2001 by the Society of Gynecologic Oncologists (SGO), it had been observed that among gynaecologic oncologists, diaphragmatic disease was reported as the second most common cause of suboptimal debulking and this was reportedly attributed to lack of training along with limited experience of the surgeons [16]. This scenario has changed over the years as elaborated in the literature in years that have followed.

In 2010, in a study of 89 patients with stage IIIB–IV epithelial ovarian malignancies recorded 4 types of diaphragmatic surgeries during interval debulking procedures: coagulation of the diaphragmatic disease, diaphragmatic stripping, combination of both these techniques, and resection of the tumour with the adjacent infiltrated part of the diaphragm muscle along with the pleura above it. Debulking to no residual tumour was achieved in majority of their patients, in 95 to 100% in the coagulation and stripping groups while it was 50% in the resection group [17]. Further studies [18] have discussed a classification of diaphragmatic surgeries into three types depending on the site and extent of surgery which was reflected in liver mobilization, exposure of the central tendon of diaphragm, bare area of the liver, inferior vena cava, and hepatic vein. To reach the posterior, most aspects of the diaphragm and exposure of the hepatic veins and IVC would require advanced surgical skills which could be used as a measure to assess the surgical training under the speciality. The most advanced surgery (type III of Tozzi’s classification) includes resection of the diseased peritoneum with diaphragmatic muscle, closure of defect with prolene, under Valsalva manoeuvre with a Foleys suction catheter in situ to create negative pressure. This requires good surgical skills which come with experience.

Pulmonary complications constitute the main morbidity of diaphragmatic surgeries and stress on respiratory status of patients with diaphragmatic perforation is the pivotal factor to be addressed in the post-operative period in order to reduce chances of dyspnoea, poor respiratory effort, and lung infection [19, 20]. Post-operative detection of these complications along with active chest physiotherapy and incentive spirometery can help to alleviate majority of these serious complications. Pre-operative disease load and other factors like large volume ascites prior to surgery, widespread intra-abdominal disease requiring visceral peritonectomy (with increased peritoneal carcinomatosis index), prolonged exposure of the bare area of the diaphragm after mobilization of the liver, release of VEGF and various inflammatory mediators, reduced intra-thoracic negative pressure secondary to post-operative pain, or phrenic nerve paralysis are some of the factors which predispose the patient to have complications [21–23].

Pleural effusion is the most frequent complication and chest tube placement or thoracocentesis may be necessary for the relief of respiratory distress [17]. It was more commonly associated with hepatic mobilization which was often the initial step to perform a safe and complete surgery in the diaphragmatic region. Routine intra-operative chest tube insertion in patients with stage IIIC–IV epithelial ovarian, fallopian tube, or peritoneal cancer who underwent debulking surgery with diaphragmatic procedure was not indicated as it did not contribute significantly to the post-operative prevention of complications [24]. Pulmonary complications were more commonly seen in the patients undergoing diaphragmatic resection versus stripping. Though the operative time was higher in the resection group, other post-operative complication rates such as pneumothorax and subphrenic abscess were comparable between both the groups [25]. In another recent study on complications of diaphragmatic surgeries [26], complications such as pneumothorax were found in 7.3% of cases and pleural effusion occurred in 33% of cases and post-operative pneumonia in 15.3%. Routine intra-operative chest tube was advocated when full thickness resection of diaphragm was done in this series.In our study, we did not advocate elective intra-operative use of intercostal chest tube except for one patient who underwent an intra-operative insertion due to pleural tear. Pleural effusion was the commonest complication in our series (32%) and pneumothorax was 1.2%. Most of them were managed conservatively, while some required a pleural tap. Majority of the ICD insertions in our series, 8/10 (80%), were in the immediate post-operative period. Ascites was present in nearly all post-op patients who had later developed pleural effusions. The risk of post-operative pleural effusion was associated with the size of the diaphragmatic resection. Intra-operative placement of a chest tube had been recommended in patients undergoing complete liver mobilization and larger diaphragmatic resections with full thickness as per some studies [10, 27, 28]. All the patients undergoing extensive diaphragmatic cytoreductive procedures, for example diaphragmatic peritonectomy and/or full thickness resection, were considered to be at high risk for developing pleural effusion. Studies have shown that pigtail catheters were safer than the traditional chest tubes in draining the effusions as they did not cause pressure over the neurovascular bundle causing less pain [29]. Though there have been reports in the literature about the role of pigtail insertion in the management of pleural effusion, in our study, we had used single lumen pleurex catheters and chest tubes for pleural fluid drainage. In our study, usage of pigtail catheters was limited to draining intra-abdominal collections.

Splenectomy was often a necessary procedure to clear the disease from left upper quadrant and was more commonly associated with a full thickness resection of the diaphragm [29]. This was not observed in our series as both diaphragmatic stripping and resection groups had equal distribution of splenectomy procedures. The right diaphragm was found to be involved in more cases in keeping with the literature [30]. Though the rates of minor complications like UTI were lesser in comparison to other studies, the grade 3–5 Clavien-Dindo complications were higher compared to other studies [31]. Mean duration of ICU stay and mean number of days spent in hospital were also comparable to literature [17, 31, 32]. Our high optimal debulking rates of 89% could be attributed to the careful selection of cases backed by a good-quality pre-operative imaging. Diaphragmatic resection was done in a few patients who eventually had suboptimal surgeries due to unavoidable circumstances such as non-consenting for stoma or unstable condition of patient during the course of surgery. Studies have suggested that if the patient cannot undergo near optimal cytoreduction, radical cytoreductive procedures should not be performed except for palliation [33]. Such extensive procedures should only be undertaken if optimal cytoreduction can be achieved; otherwise, risks would outweigh the benefits. There were no 30-day mortalities in our series.

A study assessing outcomes and morbidity following diaphragmatic peritonectomy for women with ovarian carcinoma found that in patients undergoing primary cytoreduction, median survival was 59 months and 5-year survival was 49% while for recurrent ovarian carcinoma, median survival was 23 months and 5-year survival was 16% [34]. Our median disease-free and overall survival of 32 and 22 months were comparable to available literature [28]. Studies have proved an overall survival benefit for patients undergoing diaphragmatic surgeries than those who did not in optimally debulked advanced stage primary ovarian malignancies [9, 35]. Hence, all efforts should be directed towards achieving optimal cytoreduction in ovarian cancer surgeries and diaphragmatic debulking needs to be done if required.

Conclusion

The goal of cytoreductive surgery in advanced ovarian cancer is to remove all gross tumour and this is possible only if the liver is mobilised and diaphragm stripped or resected. Careful selection of cases, surgical expertise, and good post-operative care can decrease the peri- and post-operative complications. Diaphragmatic disease clearance is an integral part of surgical repertoire of the gynaecologic oncologists. Anticipation, prevention, and management of the subsequent complications must be emphasised.

Declarations

Conflict of Interest

The authors declare no competing interests.