Abdominal computed tomography scans in the selection of patients with malignant peritoneal mesothelioma for comprehensive treatment with cytoreductive surgery and perioperative intraperitoneal chemotherapy

Authors


Abstract

BACKGROUND

Malignant peritoneal mesothelioma is a rare and fatal disease. Until recently, the treatment options were very limited and ineffective. The new comprehensive approach of cytoreductive surgery with perioperative intraperitoneal chemotherapy offers improved survival rates at a cost of considerable morbidity and mortality as in other peritoneal surface malignancies. The outcome after these treatments is predominantly dependent on adequate cytoreduction. The aim of the current study was to identify computed tomography (CT) scan images that are useful in patient selection for this comprehensive approach.

METHODS

An analysis of the preoperative CT scans of 30 patients with peritoneal mesothelioma treated with cytoreductive surgery and perioperative intraperitoneal chemotherapy at a single institution was performed. Based on the size of residual tumor nodules after cytoreductive surgery, patients were divided into 2 groups: those with residual lesions ≤ 2.5 cm (adequate cytoreduction) and those with residual lesions > 2.5 cm (suboptimal cytoreduction). The CT scans for each patient were evaluated by a standardized scoring system with the reader blinded to the operative findings. Thirty-nine CT scan parameters were obtained and statistically analyzed to determine their association with the study outcome variables, namely, adequacy of cytoreduction.

RESULTS

Seven patients (64%) in the suboptimal cytoreduction group and 2 patients (11%) in the adequate cytoreduction group had a > 5-cm tumor mass in the epigastric region (P = 0.004). Nine patients (82%) in the suboptimal group and 2 patients (11%) in the adequate cytoreduction group had CT scans that showed loss of normal architecture of the small bowel and its mesentery (P < 0.001). In a composite analysis of these 2 radiologic features, none of the patients with a > 5-cm tumor mass in the epigastric region and loss of normal architecture of the small bowel and its mesentery had an adequate cytoreduction. Patients who lacked these two preoperative CT scan findings had a 94% probability of an adequate cytoreduction.

CONCLUSIONS

CT scans effectively identified large peritoneal mesothelioma tumors at crucial anatomic sites. Because adequate cytoreduction is necessary to achieve prolonged survival, CT scans became an accurate prognostic radiologic test for patient selection for comprehensive treatment. Cancer 2005. © 2005 American Cancer Society.

Until recently, peritoneal mesothelioma was a lethal disease process with survival limited to approximately 1 year.1–7 Reports from two peritoneal surface malignancy treatment centers suggest that combined treatment with cytoreductive surgery and aggressive perioperative intraperitoneal chemotherapy greatly prolonged the survival of these patients. Park et al.8 reported that 18 patients had a median survival period of 26 months and Sebbag et al.9 reported that 33 patients had a survival period of 31 months. Recent updates increased the median survival periods to 97 and 67 months, respectively.10, 11 These clinical data suggest that a change in the treatment strategy for this rare disease has resulted in a survival benefit.

The selection factors associated with prolonged survival after comprehensive treatment have been reviewed.11 One of the most prognostically significant predictors for survival was the adequacy of cytoreduction. Patients who received an adequate cytoreduction, which resulted in residual peritoneal tumors ≤ 2.5 cm, had improved survival compared with patients who received a suboptimal cytoreduction, which resulted in residual peritoneal tumors > 2.5 cm.11 However, the size of the residual peritoneal tumors was determined after the completion of surgery. This information could not be used preoperatively in the patient selection process. To date, the radiologic findings associated with a favorable outcome with this comprehensive treatment plan have not been addressed, although computed tomographic (CT) scan descriptions of this condition have been published.12–14 CT scan criteria for selection of patients for cytoreduction and perioperative intraperitoneal chemotherapy with peritoneal surface malignancies have been successful.15

The purpose of the current study was to use preoperative abdominal and pelvic CT scans to assist in the identification of patients most likely to benefit from comprehensive treatment using cytoreductive surgery and perioperative intraperitoneal chemotherapy.

MATERIALS AND METHODS

Patient Selection

Between 1989 and 2003, 68 patients with malignant peritoneal mesothelioma were treated with cytoreductive surgery and perioperative intraperitoneal chemotherapy by the same surgeon at the Washington Cancer Institute (Washington, DC). Thirty patients in this study received a diagnosis based on tissue biopsy specimens obtained after laparotomy (n = 2), laparoscopy (n = 25), or after a CT scan-guided or ultrasound-guided technique (n = 3). In a previous study,9 we reported different pathologic types of peritoneal mesothelioma including multicystic, low-grade papillary mesothelioma, papillary, epithelial, biphasic, sarcomatous, and deciduoid mesothelioma. Patients who were diagnosed with multicystic peritoneal mesothelioma or low-grade papillary mesothelioma were excluded from the study, because they may represent diseases of a different clinicopathologic entity from other types of peritoneal mesotheliomas. Also not included in our study are patients with malignant peritoneal mesothelioma who did not have CT scans available. Among the remaining patients, 30 had a preoperative CT scan available for review and they formed the basis of the current report. All of these patients underwent an extensive preoperative workup that included a clinical examination and abdominal, pelvic, and chest CT scans. Measurements of relevant tumor markers (e.g., CA 15-3, CA 19-9, CA 72-4, CA 125, and carcinoembryonic antigen) were also obtained. If patients had no evidence of metastatic disease (except for a small amount of pleural fluid) and had a good performance status, they were offered combined treatment including cytoreductive surgery with perioperative intraperitoneal chemotherapy.11 In these patients, the median interval between the receipt of a CT scan and surgery was 2 days (range, 1–19 days). The CT scans and combined treatment were performed in the same medical center.

Computed Tomography Scans

Because the current study was performed over a 14-year time period, a variety of CT scanners were utilized. For all patients, there was a minimum 1-cm continuous slice thickness through the entire abdomen and pelvis. In addition, the radiologic studies were performed after the administration of oral and intravenous contrast media. All CT scans were performed after an angiodynamic bolus at a rate of 1–2 mL per second, for a total volume of 150–180 mL of iothalamate meglumine 60% (Conray 60, Mallinckrodt, St. Louis, MO) or iohexol 240 (Omnipaque, Sanofi Winthrop, New York, NY). A 30–60-second delay occurred between the initiation of contrast administration and the start of CT scanning. Precontrast bowel preparation included 900–1200 mL of oral barium. Oral contrast ingestion started ≥ 12 hours before the CT scan (2.1% barium sulfate suspension, E-Z-EM, Inc., Westbury, NY). Rectal barium contrast was administered immediately before the CT scan. For clinical use, the CT scans were read by a staff radiologist. For the current clinical research study, all CT scans were reread by a single physician to score abdominal and pelvic tumor deposits using a standardized scoring system. The physician was aware that all patients in the study had clinical evidence of peritoneal mesothelioma and that some anatomic sites present greater difficulty for adequate cytoreduction, but was masked to the operative findings.

CT scan assessment of peritoneal disease was evaluated using a peritoneal cancer index (PCI). First, the tumor size was evaluated in 13 abdominopelvic regions: the umbilical region, right upper quadrant, epigastrium, left upper quadrant, left flank, left lower quadrant, pelvis, right lower quadrant, right flank, upper jejunum, lower jejunum, upper ileum, and the lower ileum (Fig. 1). For the PCI data accumulation, it was assumed that the upper and lower jejunum are in the left upper and left lower quadrants, respectively, and that the upper and lower ileum are in the right upper and right lower quadrants, respectively. Then, the tumor size was assessed specifically at 16 abdominopelvic anatomic sites: the abdominal wall, greater omentum, right hemidiaphragm, left hemidiaphragm, liver, spleen, stomach/duodenum, lesser omentum, subpyloric space, pancreas, small bowel and its mesentery, large bowel and its mesentery, right paracolic gutter, left paracolic gutter, pouch of Douglas, and the retroperitoneum.

Figure 1.

Abdominopelvic regions 0–12.

In the 13 abdominopelvic regions and 16 abdominopelvic anatomic sites, CT scan assessment of tumor size was categorized into 4 groups: 0, no detectable disease; 1, minimal disease (tumor thickness < 0.5 cm); 2, moderate disease (tumor thickness ≥ 0.5 cm and ≤ 5 cm); and 3, macroscopic disease (tumor thickness > 5 cm).

Ten potentially prognostic interpretative CT scan parameters were selected and assessed. These were the presence of a scalloping appearance of the liver surface, the presence of small bowel obstruction, the presence of large bowel obstruction, the presence of biliary tract obstruction, the presence of ureteral obstruction, the presence of loculated ascites versus free ascites versus no ascites, the presence of greater omental caking, the presence of pleural effusion, the presence of pleural thickening, and a descriptive CT scan classification (Class 0–III) of the small bowel and its mesentery (Table 1).

Table 1. Intepretative CT Scan Classification of Small Bowel and its Mesentery
ClassPresence of ascitesSmall bowel and mesentery involvementLoss of mesenteric vessel clarityCT scan interpretation
  1. CT: computed tomography.

0NoNoNoNormal appearance
IYesNoNoAscites only
IIYesThickening, enhancingNoSolid tumor present
IIIYesNodular thickening, segmental obstructionYesLoss of normal architecture

As indicated in Table 1, characteristic interpretative CT appearances of the small bowel and its mesentery were categorized into four classes. First, Class 0 CT scans showed no ascites in the region of the small bowel. The jejunal and ileal vessels appeared as round and curvilinear densities within the mesenteric fat. Second, Class I CT scans showed free intraperitoneal fluid only. The mesentery was stranded and stratified as the fluid accumulation outlined the small bowel mesentery. The small bowel vessels are identified easily within the mesenteric fat (Fig. 2). Third, Class II CT scans showed tumor involvement of the small bowel and/or its mesentery. The peritoneal surface was thickened and enhanced due to the presence of tumor nodules (usually half-spherical bodies) or plaques (flat implants whose diameter was greater than thickness). There may be an increased amount of ascitic fluid and the mesentery may appear to be stellate or pleated. The small bowel mesenteric vessels were still identifiable (Fig. 3). Fourth, Class III CT scans showed increased solid tumor involvement and adjacent small bowel loops are matted together in some cuts. The configuration of the small bowel and its mesentery was distorted and thickened. Segmental small bowel obstruction is present. Intraperitoneal fluid may be loculated. The small bowel mesenteric vessels were difficult to define on some cuts due to obliteration of mesenteric fat (Fig. 4).

Figure 2.

Computed tomography scan shows free intraperitoneal fluid only. The mesentery becomes stranded and stratified as the fluid accumulation outlines the small bowel mesentery. The small bowel vessels are identified easily within the mesenteric fat.

Figure 3.

Computed tomography scan shows tumor involvement of the small bowel mesentery. The peritoneal lining is thickened and enhanced due to the presence of tumor nodules or plaques. There may be an increased amount of ascitic fluid. The small bowel mesenteric vessels are still identifiable.

Figure 4.

Computed tomography scan shows increased solid tumor involvement with adjacent bowel loops matted together. The configuration of the small bowel and its mesentery is distorted and thickened. Intraperitoneal fluid may be loculated. The small bowel mesenteric vessels are difficult to define due to the obliteration of mesenteric fat density.

Cytoreductive Surgery

All patients underwent cytoreductive surgery by the same surgeon (PHS) with an intention to remove all visible intraperitoneal tumor deposits. After a midline abdominal incision, a thorough exploration of the abdomen and pelvis was performed. The location, tumor size, and extent of intraabdominal tumor deposits were recorded prospectively in the operative notes. The details of the surgical techniques are described elsewhere.16 In short, cytoreductive surgery consists of one to five peritonectomy procedures including greater omentectomy with splenectomy, left upper quadrant peritonectomy, right upper quadrant peritonectomy, lesser omentectomy with cholecyctectomy, and pelvic peritonectomy with rectosigmoid colon resection. All positive findings were documented by pathologic examination of each of the resected specimens. All sites and residual tumors after cytoreductive surgery also were recorded prospectively. Patients who had residual tumor nodules ≤ 2.5 cm at all anatomic sites within the abdomen or pelvis were scored as having received adequate cytoreduction. Patients who had residual tumors > 2.5 cm at even a single location were regarded as receiving suboptimal cytoreduction. This categorization was based on the previous clinical research of 68 patients with primary peritoneal mesothelioma.11

Perioperative Intraperitoneal Chemotherapy

After maximal cytoreduction of the peritoneal mesothelioma, heated intraoperative intraperitoneal chemotherapy with cisplatin (50 mg/m2) and doxorubicin (15 mg/m2) was administered at approximately 41.5 °C in 3 L of 1.5% dextrose peritoneal dialysis solution for 90 minutes. The intraperitoneal chemotherapy solution was distributed manually to facilitate maximal contact and penetration of the chemotherapy solution into residual cancer cells. This was a planned part of the surgical procedure. In the early postoperative period, 20 mg/m2 per day paclitaxel on postoperative Days 1–5 was administered into the peritoneal cavity in 1 L of 1.5% dextrose peritoneal dialysis solution or in 1 L of 6% hetastarch solution.11

Statistical Analysis

The CT scan and the cytoreductive surgery were performed as routine clinical management of this disease. Two-thirds of the patients participated in an institutional review board (IRB)-approved pharmacologic study. The IRB granted permission to review the clinical data of these 30 patients. The distribution of the different CT scan parameters between the two groups of patients was analyzed by Fisher exact test using the Statistical Package for Social Sciences for Windows (Version 9.0; SPSS GmbH, Munich, Germany). A significant difference was defined as P < 0.01. All CT scan parameters that were associated significantly with cytoreductive surgery outcome in the univariate analysis were included in a tree-structured diagram, which was constructed to provide an estimated clinical pathway based on the CT scan criteria.17 All combinations of the selected CT scan parameters were compared to obtain the best prediction of cytoreductive surgery outcome.

RESULTS

The clinical characteristics of the 30 patients with CT scans available for study are shown in Table 2. Nineteen of the patients received adequate cytoreduction and 11 received suboptimal cytoreduction.

Table 2. Clinical Characteristics of Patients with Adequate Versus Suboptimal Cytoreductiona
Characteristics No.Adequate cytoreduction (%)Suboptimal cytoreduction (%)
  • CT: computed tomography.

  • a

    There were no statistically significant differences in clinical characteristics of patients with adequate and suboptimal cytoreduction.

No. of patients1911
Median age at diagnosis (range) (yrs)50 (28–72 yrs)52 (23–76 yrs)
Median age at surgery (range) (yrs)50 (29–72 yrs)52 (23–74 yrs)
Gender  
 Male12 (63) 8 (73)
 Female 7 (37) 3 (27)
Median time to diagnosis and operation interval (range) (mos) 3 (1–11 mos) 4 (0–25 mos)
Median time to CT scans and operation interval (range) (days) 2 (1–19 days) 2 (1–6 days)
Pathology (%)  
 Papillary/epithelial13 (68%) 9 (82%)
 Sarcomatoid/deciduoid/biphasic 6 (32%) 2 (18%)

Analysis of Computed Tomography Scan Findings in 13 Abdominopelvic Regions by Adequate Versus Suboptimal Cytoreduction

Tumor size as assessed by CT scans of the 13 abdominopelvic regions for patients with adequate or suboptimal cytoreduction is shown in Table 3. The tumor size in abdominopelvic region 2, the epigastric region, was statistically and significantly different in these 2 groups of patients (P = 0.002). Sixty-four percent of the patients in the suboptimal cytoreduction group versus 11% in the adequate cytoreduction group had a > 5-cm tumor mass in the epigastric region (Fig. 5). Also shown in Table 3 are the data regarding tumor size in the four small bowel regions. In the upper and lower jejunum and in the upper ileum there was a statistically and significant larger tumor mass in patients with suboptimal cytoreduction compared with patients with an adequate cytoreduction. For the small bowel regions, all analyses except for the distal ileum were significant. Other abdominopelvic regions did not show any significant differences between the two groups.

Table 3. Analysis of CT Scan Findings in 13 Abdominopelvic Regions by Adequate Versus Suboptimal Cytoreduction.a
Tumor sizeaAdequate cytoreductionSuboptimal cytoreductionP valuebTumor sizeaAdequate cytoreductionSuboptimal cytoreductionP valueb
No. (n = 19) (%)No. (n = 11) (%)No. (n = 19) (%)No. (n = 11) (%)
  • CT: Computed tomography; NS: not significant.

  • a

    0: No tumor; 1: tumor diameter < 0.5 cm; 2: tumor diameter is 0.5–5 cm; 3: tumor diameter > 5 cm.

  • b

    NS (Fisher exact test, P < 0.01 was assumed to be significant).

0 (Central)   7 (Right lower)   
 —— (—)— (—)NS —— (—)— (—)NS
 02 (11)0 (0)  02 (11)0 (0) 
 11 (5)0 (0)  13 (16)0 (0) 
 24 (21)1 (9)  211 (58)5 (45) 
 312 (63)10 (91)  33 (16)6 (55) 
1 (Right upper)   8 (Right flank)   
 —— (—)— (—)NS —— (—)— (—)NS
 04 (21)0 (0)  05 (26)0 (0) 
 16 (32)0 (0)  12 (11)0 (0) 
 29 (47)9 (82)  211 (58)7 (64) 
 30 (0)2 (18)  31 (5)4 (36) 
2 (Epigastric)   9 (Upper jejunum)   
 —— (—)— (—)0.002 —— (—)— (—)0.005
 09 (47)0 (0)  05 (26)0 (0) 
 14 (21)0 (0)  17 (37)0 (0) 
 24 (21)4 (36)  25 (26)10 (91) 
 32 (11)7 (64)  32 (11)1 (9) 
3 (Left upper)   10 (Lower jejunum)   
 —— (—)— (—)NS —— (—)— (—)0.008
 04 (21)0 (0)  05 (26)0 (0) 
 17 (37)0 (0)  17 (37)0 (0) 
 26 (32)6 (55)  25 (26)9 (82) 
 32 (11)5 (45)  32 (11)2 (18) 
4 (Left flank)   11 (Upper ileum)   
 —— (—)— (—)NS —— (—)— (—)0.004
 05 (26)0 (0)  04 (21)0 (0) 
 15 (26)0 (0)  19 (47)0 (0) 
 29 (47)9 (82)  25 (26)10 (91) 
 30 (0)2 (18)  31 (5)1 (9) 
5 (Left lower)   12 (Lower ileum)   
 —— (—)— (—)NS —— (—)— (—)NS
 03 (16)0 (0)  04 (21)0 (0) 
 12 (11)0 (0)  17 (37)0 (0) 
 212 (63)4 (36)  25 (26)9 (82) 
 32 (11)7 (64)  33 (16)2 (18) 
6 (Pelvis)       
 —— (—)— (—)NS    
 01 (5)0 (0)     
 11 (5)0 (0)     
 25 (26)1 (9)     
 312 (63)10 (91)     
Figure 5.

Computed tomography scan shows lesion size 3 in the epigastric region. Solid tumor infiltrates the lesser omentum and fills the lesser omentum and lesser sac.

Analysis of Computed Tomography Scan Findings in 16 Anatomic Sites by Adequate Versus Suboptimal Cytoreduction

The relation of 16 anatomic sites to the results of cytoreduction is shown in Table 4. Four sites were statistically significant, including the stomach/duodenum (P < 0.001), the lesser sac (P = 0.002), the subpyloric space (P < 0.001), and the small bowel with its mesentery (P = 0.005).

Table 4. Analysis of CT Scan Findings in 16 Anatomic Sites by Adequate Versus Suboptimal Cytoreductiona
Tumor sizeaAdequate cytoreductionSuboptimal cytoreductionP valueTumor sizeaAdequate cytoreductionSuboptimal cytoreductionP value
No. (n = 19) (%)No. (n = 11) (%)No. (n = 19) (%)No. (n = 11) (%)
  • CT: Computed tomography; NS: not significant.

  • a

    0: No tumor; 1: tumor diameter < 0.5 cm; 2: tumor diameter is 0.5–5 cm; 3: tumor diameter > 5 cm.

  • b NS (Fisher exact test, P < 0.01 was assumed to be significant).

Abdominal wall   Subpyloric space   
 —— (—)— (—)NS —— (—)— (—)< 0.001
 03 (16)0 (0)NS 010 (53)0 (0) 
 15 (26)0 (0)  15 (26)0 (0) 
 29 (47)10 (91)  24 (21)7 (64) 
 32 (11)1 (9)  30 (0)4 (36) 
Greater omentum   Pancreatic surface   
 —— (—)— (—)NS —— (—)— (—)NS
 02 (11)0 (0)  018 (95)6 (55) 
 11 (5)0 (0)  11 (5)2 (18) 
 24 (21)0 (0)  20 (0)3 (27) 
 312 (63)11 (100)  30 (0)0 (0) 
Right hemidiaphragm   Small bowel/mesentery   
 —— (—)— (—)NS —— (—)— (—)0.005
 07 (37)0 (0)  04 (21)0 (0) 
 13 (16)0 (0)  18 (42)0 (0) 
 29 (47)9 (82)  24 (21)9 (82) 
 30 (0)2 (18)  33 (16)2 (18) 
Left hemidiaphragm   Right paracolic gutter   
 —— (—)— (—)NS —— (—)— (—)NS
 010 (53)0 (0)  03 (16)0 (0) 
 11 (5)0 (0)  14 (21)0 (0) 
 26 (32)9 (82)  29 (47)5 (45) 
 32 (11)2 (18)  33 (16)6 (55) 
Hepatic capsule   Left paracolic gutter   
 —— (—)— (—)NS —— (—)— (—)NS
 08 (42)0 (0)  04 (21)0 (0) 
 13 (16)1 (9)  13 (16)0 (0) 
 28 (42)10 (91)  210 (53)5 (45) 
 30 (0)0 (0)  32 (11)6 (55) 
Splenic capsule   Large bowel/mesentery   
 —— (—)— (—)NS —— (—)— (—)NS
 010 (53)0 (0)  04 (21)0 (0) 
 11 (5)1 (9)  11 (5)0 (0) 
 28 (42)10 (91)  24 (21)4 (36) 
 30 (0)0 (0)  310 (53)7 (64) 
Stomach/duodenum   Pouch of Douglas   
 —— (—)— (—)< 0.001 —— (—)— (—)NS
 013 (68)0 (0)  01 (5)0 (0) 
 14 (21)1 (9)  10 (0)0 (0) 
 22 (11)10 (91)  28 (42)2 (18) 
 30 (0)0 (0)  310 (53)9 (82) 
Lesser sac   Retroperitoneal space   
 —— (—)— (—)0.002 —— (—)— (—)N/A
 09 (47)0 (0)  00 (0)0 (0) 
 14 (21)0 (0)  10 (0)0 (0) 
 24 (21)4 (36)  20 (0)0 (0) 
 32 (11)7 (64)  30 (0)0 (0) 

Analysis of Computed Tomography Scan Interpretative Findings by Adequate Versus Suboptimal Cytoreduction

The presence of pleural effusion and the classification of the interpretive appearances of the small bowel and its mesentery were statistically significant, with P values of 0.003 and 0.001, respectively (Table 5). Nine of 11 patients (82%) with the suboptimal cytoreduction and 2 of 19 patients (11%) with adequate cytoreduction had Class III interpretative findings (P < 0.001).

Table 5. Analysis of CT Scan Interpretative Findings by Adequate Versus Suboptimal Cytoreduction
CT scan parametersAdequate cytoreductionSuboptimal cytoreductionP valuea
No. (n = 19) (%)No. (n = 11) (%)
  • CT: computed tomography; NS: not significant; N/A: not applicable.

  • a

    NS (Fisher exact test, P < 0.01 was assumed to be significant).

Liver surface scalloping— (—)— (—)NS
 Present17 (89)6 (55) 
 Absent2 (11)5 (45) 
Small bowel obstruction— (—)— (—)NS
 Present1 (5)3 (27) 
 Absent18 (95)8 (73) 
Large bowel obstruction— (—)— (—)NS
 Present3 (16)6 (55) 
 Absent16 (84)5 (45) 
Biliary tract obstruction— (—)— (—)N/A
 Present0 (0)0 (0) 
 Absent19 (100)11 (100) 
Ureteral obstruction— (—)— (—)NS
 Present0 (0)3 (27) 
 Absent19 (100)8 (73) 
Ascites— (—)— (—)NS
 Present—free layering9 (47)4 (36) 
 Present—loculated2 (11)5 (45) 
 Absent8 (42)2 (18) 
Greater omental caking— (—)— (—)NS
 Present11 (58)9 (82) 
 Absent8 (42)2 (18) 
Pleural effusion— (—)— (—)0.003
 Present0 (0)5 (45) 
 Absent19 (100)6 (55) 
Pleural thickening— (—)— (—)NS
 Present3 (16)4 (36) 
 Absent16 (84)7 (64) 
CT scan classification of small bowel and its mesentery— (—)— (—)< 0.001
 Group 02 (11)0 (0) 
 Group I8 (42)0 (0) 
 Group II7 (37)2 (18) 
 Group III2 (11)9 (82) 

Analysis Using a Tree-Structured Diagram

Statistical analysis using a tree-structured diagram was performed to provide a model for a clinical pathway determined by preoperative CT scans for treating patients with peritoneal mesothelioma. The CT scan parameters tested in the tree were abdominopelvic region 2, the first three regions of the small bowel, the stomach/duodenum, lesser sac, subpyloric space, the small bowel and its mesentery, the presence of pleural effusion, and the CT scan interpretative classification of the small bowel with its mesentery. Lesion size III in abdominopelvic region 2, the epigastrium, was determined to improve the predictive value of the interpretative CT scan classification of the small bowel with its mesentery. According to this model, a patient with a tumor size > 5 cm in the epigastric region and with a Class III appearance of the small bowel and its mesentery has a probability of 100% for a suboptimal cytoreduction. A patient without either of these 2 preoperative CT scan findings has a 94% probability of an adequate cytoreduction. Adding the other CT scan parameters did not improve the predictive value of the tree (Fig. 6).

Figure 6.

The predictive value of Computed tomography scan findings by a tree-structured diagram. AC: adequate cytoreduction; SC: suboptimal cytoreduction; SB/SBM: small bowel and small bowel mesentery.

DISCUSSION

Between 1989 and 2003, 68 patients with primary peritoneal mesothelioma underwent this comprehensive treatment strategy of cytoreductive surgery with perioperative intraperitoneal chemotherapy at the Washington Cancer Institute. The overall median survival period was 67 months from the time of cytoreductive surgery. In the same series, a retrospective analysis of prospectively collected clinical information was performed to determine the significant quantitative prognostic indicators for survival. One of the most prognostically significant predictors of survival was the size of residual tumor nodules, which was determined by the surgeon at the completion of the procedure. Residual peritoneal tumors > 2.5 cm predicted poor survival compared with residual peritoneal tumors ≤ 2.5 cm.11 In these data, the size of the residual tumor nodules could only be determined after the completion of surgery and could not be used in the patient selection process. No data from our patients regarding survival in the absence of surgery were available. The literature would suggest a survival period of < 1 year.1–7

The current study indicates that the preoperative CT scan provides information that can assist in the selection of patients for this comprehensive treatment. The anatomic location of a large tumor is a determinant of the surgical outcome. The anatomy of some parts of the abdomen and pelvis allows for complete or near complete removal of large mesothelioma tumors. However, large tumors at other anatomic sites prohibit adequate cytoreduction. Surgically, two difficult regions for adequate cytoreduction of large tumors are the epigastric region and the small bowel. The CT scan in identifying large tumors at these two anatomic sites was of great assistance in predicting an adequate versus suboptimal cytoreduction.

In performing cytoreductive surgery for malignant peritoneal mesothelioma, the greater omentum is usually the most extensively involved intraabdominal structure. It is removed along with the gastroepiploic vascular arcade. Usually, this is one of the early procedures in the cytoreduction. The sole remaining blood supply to the stomach is the lesser omental vascular arcade composed of the right and left gastric arteries. If this vasculature is compromised, devascularization of the body of the stomach will occur. Large tumors in the epigastric region will extensively involve the right gastric/left gastric vascular arcade. Unless the surgeon is willing to perform a total gastrectomy, this tumor must not be resected. Consequently, a large solid tumor in the epigastric region enables suboptimal cytoreduction.

In the performance of adequate cytoreduction with perioperative intraperitoneal chemotherapy, the small bowel and its mesentery are dissected. The large tumor is removed and all bowel loops and mesenteric leaves are separated to allow contact between the residual tumor and perioperative intraperitoneal chemotherapy. If adjacent intestinal and mesenteric structures are matted together, adequate tumor removal and locoregional chemotherapy are not possible. Class III CT scan findings of the small bowel and its mesentery were associated with inadequate surgical and regional chemotherapy treatment.

The adequacy of the combined treatment was predicted accurately by the CT scan identification of not only a large tumor but also its anatomic location. Tumor measurements alone were not sufficient. For example, large tumors in these patients were routinely present in the pelvis. However, with the complete pelvic peritonectomy, tumors at this site were removed completely. Therefore, large tumors in abdominopelvic region 6 had no impact on survival and a preoperative CT scan had no prognostic implications. The same rationale applies to the extensive infiltration of the greater omentum by mesothelioma referred to as omental caking. Tumors in the greater omentum, no matter the size, can be resected.

Pseudomyxoma peritonei is another peritoneal surface malignancy successfully treated by cytoreductive surgery and perioperative intraperitoneal chemotherapy. Jacquet et al.15 studied the prognostic implications of a preoperative CT scan. They concluded that CT scan findings for identification of patients who would have an incomplete cytoreduction were segmental obstruction of the small bowel and large tumor nodules on the small bowel and its mesentery. In contrast to patients with malignant peritoneal mesothelioma, patients with pseudomyxoma peritonei who had large tumors in the epigastric region (abdominopelvic region 2) involving the vasculature of the stomach did not have a poor prognosis. The patients who had extensive perigastric disease underwent a total gastrectomy and their survival results are comparable to other patients who received a complete cytoreduction.18 In peritoneal mesothelioma, total gastrectomy has not been added to the list of peritonectomies that are performed in these patients. Therefore, extensive disease as assessed by a lesion size score of 3 in the epigastric region was associated with inadequate cytoreduction and a poor prognosis. As the success of locoregional treatments for malignant mesothelioma begins to compare with those for pseudomyxoma peritonei, a further surgical evolution to total gastrectomy for these patients may be considered.

A single radiologist (NH) was responsible for the interpretative classification of the small bowel and its mesentery. This may be a limitation of the quality of the data provided in the current study. However, as the surgeons and radiologist read and reread these CT scans, it was accepted that the Class 0–III criteria represented a spectrum of disease severity. As our experience with the findings seen within the small bowel and its mesentery increased, it became apparent that Class I and II were not always readily distinguishable, but that Class 0, I, and II were always distinguishable from Class III. In none of the patients was there difficulty in separating Class 0, I, and II from Class III, by using the three criteria shown in Table 1.

A surgical finding that is of great importance in the selection of patients for adequate cytoreduction is the presence versus the absence of foreshortening of the small bowel mesentery. Although we searched for radiologic evidence of small bowel retraction in these CT scans, we were not able to find reliable radiologic criteria for this surgical finding. In some patients, foreshortening of the small bowel mesentery was difficult or impossible to appreciate on the CT scan.

These data regarding CT scan findings are only meaningful for patients treated with this comprehensive approach utilizing cytoreductive surgery and perioperative intraperitoneal chemotherapy. These data used in patients managed only with systemic chemotherapy may not be valid. Also, it is probable that this CT scan assessment of prognosis in patients managed by debulking surgery only without optimal cytoreduction plus perioperative intraperitoneal chemotherapy would have no significance. It should be noted that adequate cytoreduction was defined as tumor resection to a nodule size ≤ 2.5 cm. For patients with gastrointestinal carcinoma this would not be considered adequate cytoreduction and for patients with colon carcinoma, only cytoreduction to ≤ 2.5 mm would be considered optimal.19

In our study, CT scans were able to select for treatment those patients who had an improved long-term survival. The data are not mature enough to predict disease remission in this group of patients. However, the data do demonstrate that there is a group of patients (one-third of the total number of patients in the study) with a poor prognosis who will receive limited benefit from this combined treatment. These patients may receive surgery with intraperitoneal chemotherapy for limited goals such as tumor debulking and the alleviation of debilitating ascites. However, as systemic chemotherapy for peritoneal mesothelioma shows improved results, the CT scan can be used to select patients with a poor prognosis for aggressive neoadjuvant chemotherapy before aggressive surgery and locoregional chemotherapy. This assumes that patients with large tumors in the epigastric region and within the small bowel and its mesentery have an adequate performance status score, and can be expected to have a reasonable response rate to the neoadjuvant treatments.

A reality that accompanies cytoreductive surgery with perioperative intraperitoneal chemotherapy is its high morbidity and cost. An avoidance of unnecessary or low-value (in terms of survival) procedures would greatly improve the results of this new approach to peritoneal surface malignancy. In patients in whom CT scans suggest a poor outcome, surgery may be cancelled completely. Certainly, if the patient is a poor surgical risk, an aggressive approach should be replaced with a modified, reduced-risk procedure. In the poor prognosis group identified by CT scan criteria, the surgeon should avoid the “point of no return” in patients with a peritoneal surface malignancy. This is an enterotomy above partial or complete obstruction of the bowel. Repair has a high incidence of fistula formation. Multiple enterotomies may require an extensive resection to prevent postoperative complications but provide little or no survival benefit.

These findings may have implications for the management of other peritoneal surface malignancies. The use of CT scans to select patients with mucinous carcinoma for treatment has been described in a previous report.15 Reliable preoperative prognostic indicators are a great need in the management of patients with peritoneal surface malignancies.

We only used CT scans performed at this institution for the analysis. It is likely that technical aspects of the CT scan may be important for an optimal study. Maximal oral barium contrast that fills the entire gastrointestinal tract is beneficial. Intrarectal contrast facilitates interpretation of the pelvic CT scan. Intravenous contrast injected to maximize density differences between normal tissue, tumor layered out on peritoneal surfaces, and ascites fluid is beneficial. The time interval between intravenous contrast infusion and spiral CT scans may be crucial to determinations of ascites fluid versus solid tumor. A 60-second delay may be optimal in this situation. This distinction is important to separate the Class I, Class II, and Class III interpretative small bowel classification.

Fourteen years have elapsed since the initiation of the current study. Several different CT scanners were utilized to study these patients. Currently, helical CT scan equipment with multiple detectors is used routinely. It is likely that the information regarding CT scan assessment of prognosis will increase as more modern CT scan technology is employed routinely in these patients. Also, gadolinium-enhanced magnetic resonance imaging scans may be shown to be useful in this clinical situation.

As pointed out in the article by Sebbag et al.,9 no staging system currently exists for peritoneal mesothelioma. We believe that our work with preoperative CT scans will be of benefit in the construction of a staging system for this disease. CT scans, along with the histologic type (aggressive vs. nonaggressive) of the tumors, may be profound determinants of prognosis. However, other traditional prognostic indicators such as lymph node positivity or negativity (N factor) are unlikely to be of value. Also, the presence of disease at other sites (M factor) is extremely unusual in this disease, which is almost universally confined to the peritoneal cavity throughout its natural history. A positive peritoneal cytology so profoundly important in gastrointestinal malignancies has no meaning in these patients, who have diffuse seeding of the peritoneal surfaces. In our judgement, anatomic interpretation of preoperative CT scan findings may be an important aspect of staging when one considers the comprehensive treatment for these patients.

Ancillary