Clinical Application and Economic Implications of PET in the Assessment of Colorectal Cancer Recurrence: A Retrospective Study

Can PET's accuracy better identify resectable colorectal tumor recurrence?

INTRODUCTION
Malignant tumor cells have been found to have a high rate of glycolysis to be associated with increased membrane glucose transfer capability. (F-18) fluorodeoxyglucose (FDG) has proven useful for tracing glucose metabolism, for detecting malignant tissue and for quantifying changes in tumor glycolysis during and after treatment. Clinical applications for FDG-PET have been introduced for diagnosis, staging and metastatic survey of various types of cancers in the past five years. When physicians have used PET to address diagnosis and therapy questions, a significant positive impact has been reported.

CLINICAL QUESTION
There are approximately 110,000 new colon cancer and 45,000 new rectal cancer cases reported every year. About 70% of patients are resectable with curative intent, and about 2/3 of these patients are cured by resection. The cancer in the remaining 1/3 recurs, most commonly during the first two years after resection. Some of these patients have a recurrence limited to a regional site amenable to a surgical resection and potential cure. Distinguishing which of these patients can benefit from surgical resection with conventional diagnostic techniques is a significant clinical problem, because only about 1/4 of these patients are actually curable. The rest have the risk of being treated with unnecessary, expensive surgeries resulting in morbidity.

The clinical questions become: 1) Can FDG-PET more accurately stage the patient with potentially resectable colorectal cancer? and 2) Can FDG-PET, by decreasing the number of unnecessary procedures including surgeries, be more cost-effective in this clinical setting?

A RETROSPECTIVE VIEW OF CONVENTIONAL DIAGNOSTIC TECHNIQUES VERSUS PET
Follow-up procedures routinely used after primary colorectal cancer resection for recurrence detection are inefficient. Recent studies have shown that CEA was only 59% sensitive and 84% specific in detecting colorectal cancer recurrence. Accuracy of barium enema studies for local recurrence are in the 80% range, however, the test is only 49% sensitive and 85% specific for overall recurrence detection.

In patients with recurrent, potentially resectable disease, the pre-surgical workup goal is to determine isolated resectable disease from widespread metastases. Current diagnostic studies are not sensitive enough, so false negatives are common. CT, the benchmark staging test in colorectal cancer, has reported staging accuracies of 25-73%. CT is particularly useful for hepatic metastases, but for nodal involvement, CT sensitivity appears to be lower. Despite a negative CT, 25-50% of patients will have non-resectable lesions at the time of exploratory laparotomy.

In addition, in patients undergoing exploration for recurrent colorectal cancer, the ubiquitous presence of surgical adhesions or the limitations of surgical exposure (transverse upper abdominal incision for liver resection) often preclude a detailed operative staging. Of those resected with seemingly isolated regional disease, only 25-50% will be cured.

There is evidence that FDG-PET scanning will have a role in assessing colorectal cancer patients with suspected recurrence. In an ICP survey of 14 PET sites, the records of 267 colorectal cancer patients who had PET scans were reviewed. This group of patients included:

28 evaluated for pre-operative staging
20 evaluated for extrahepatic extent prior to hepatic resection
58 evaluated for positive CEA with a negative CT
119 evaluated for an indeterminate mass on CT
42 studied for other reasons

The initial analysis of these patients includes a subset of 54 lesions which had site-by-site surgical confirmation. In this surgical series, overall accuracy was 88% for detecting tumor vs. benign disease, and sensitivity was 93%.

STUDY APPROACH
Colorectal cancer is avid for FDG, and preliminary study supports a strong potential role for FDG-PET scanning in the imaging of the colorectal cancer patient. An ICP team was formed to collect data on the clinical and economic efficacy of PET in colorectal cancer recurrence detection after surgical resection.

CONCLUSIONS
PET can save between $5423 and $32,123 per patient by replacing laparotomy and avoiding unnecessary surgery. Also, PET examines the whole body for recurrence and is not limited to the abdomen as is laparotomy.

Based on surgical experience, inappropriate resection of recurrent cancer despite occult non-resected disease can cost up to $20,000-$30,000 per case with no improvement in outcome.

Clinical Efficacy and Cost-Effectiveness of FDG PET in Recurrent Colorectal Cancer

Presented by Peter Valk, MD, Northern California PET Imaging Center, at the 10th Annual Institute for Clinical PET Conference, Boston, MA.

Colorectal cancer is one of the few malignancies where recurrent tumor is commonly treated by surgical resection. Resection for cure may be undertaken in hepatic, pelvic or pulmonary recurrence, where the tumor is thought to be resectable in extent and localized to the anatomic site in question. However, after conventional evaluation of localized hepatic recurrence, 25-50% of patients are found to have non-resectable tumor at surgery, and the five year survival in patients who do undergo resection for cure is approximately 25%. A more accurate means than CT imaging is clearly needed for diagnosis and staging of recurrent colorectal cancer.

Diagnostic Accuracy of PET and CT: In a study of the accuracy of PET imaging (1), we imaged 134 patients with suspected colorectal cancer recurrence. Seventy-eight patients were studied for pre-operative assessment of recurrence that had been diagnosed by CT, 32 were studied because of serum CEA elevation, and 24 had other indications. Non-corrected whole-body images and static attenuation-corrected images of the region of suspected abnormality and the upper abdomen were obtained. PET findings were correlated with CT findings in 115 patients. A final diagnosis was established at 171 anatomic sites of PET and/or CT abnormality, and tumor was also found at 8 sites that were normal by imaging. Validation was by histology at 57% of sites, serial CT imaging at 25%, and clinical progression at 18%.

In the 115 patients who had been imaged by both modalities, PET sensitivity and specificity were 93% and 98% respectively, compared to 69% and 96% for CT. 95% confidence intervals for the differences between the modalities were 16-32% for sensitivity and 1-5% for specificity. Sensitivity of both modalities varied with the anatomic site of recurrence, being lowest in the abdomen (79% PET, 46% CT). In 32 patients with serum CEA elevation (18 CT negative, 14 no CT), PET sensitivity and specificity were 90% and 92%. In the patients with negative CT findings, PET was true positive in 67% (2).

These findings are similar to those reported by other investigators. Schiepers et al evaluated 83 patients with known or suspected recurrent colorectal cancer (3). They correlated PET findings with results of CT and/or ultrasound imaging of the abdomen and pelvis. PET detected 92/96 tumor sites (96%), whereas CT/US detected 56/79 sites (71%). Delbeke et al (4) found that PET was more accurate than CT for detection of liver metastases (92% vs. 78%) in 61 studies for tumor recurrence, and was also more accurate for detection of extra-hepatic tumor (92% vs. 71%). PET detected unsuspected metastases in 28% of studies. Lai and coworkers (5) demonstrated unsuspected extrahepatic metastasis in 11/34 patients (32%) with known hepatic metastasis.

Impact of PET on Patient Management:
The clinical impact of PET findings in surgical management of recurrent colorectal cancer was determined by evaluation of the post-PET treatment records of 78 patients, who had been imaged specifically for pre-operative evaluation. These patients all had evidence of a single site of recurrence that appeared to be limited and resectable by CT. In 23 patients (29%), PET demonstrated recurrence at other site(s) as well, and in two patients (3%) with hepatic recurrence, PET demonstrated extensive bilobar disease. In four patients (8%), PET correctly demonstrated no tumor at the site of CT abnormality. Forty-two of these patients subsequently underwent surgery, and non-resectable, PET-negative tumor was found in seven patients (17%). Thirty-five patients underwent curative resection.

Surgical procedures were avoided by demonstration of non-resectable tumor in 25 patients and included the following: hepatic resection 7, abdominal or pelvic tumor resection 16, and pulmonary resection 2. Laparotomy was also avoided in three patients and liver resection in one patient with false-positive CT findings. Cost savings resulting from avoidance of unnecessary surgical procedures were calculated on the basis of national Medicare reimbursement rates and a PET charge of $1,800. Net savings amounted to a total of $434,000 or $3,760 per patient.

Three previous studies have evaluated the effect of PET findings on management decisions in recurrent colorectal cancer, with findings similar to our own. Beets and coworkers (6) found that PET findings affected management in 15/35 patients with liver or pelvic recurrence (40%), and demonstrated unsuspected, non-resectable tumor in 6/24 preoperative patients (25%). In staging suspected liver metastasis only, Vitola et al. (7) found that PET altered surgical plans in 6/24 patients (25%), avoiding surgery in three patients (13%). In a similar group (5), PET avoided inappropriate hepatic resection in 8/34 patients (24%), and postponed surgery in two.

Conclusion:
Whole-body metabolic PET imaging is more accurate than CT for detection of recurrent colorectal cancer, and is a cost-effective means of differentiating resectable from non-resectable disease. PET imaging is indicated whenever a major management decision depends upon accurate evaluation of tumor presence and extent.

1. Valk PE, Abella-Columna E, Tesar RD, Pounds TR, Haseman MK, Myers RW. Detection of recurrent colorectal cancer by FDG PET in patients with serum CEA elevation. J Nucl Med 1998; 39 (Abstract book).
2. Valk PE, Abella-Columna E, Tesar RD, Pounds TR, Haseman MK, Myers RW. Comparison of PET and CT in recurrent colorectal cancer: accuracy by site of recurrence. J Nucl Med 1998; 39 (Abstract book).
3. Schiepers C, Penninck F, De Vadder N, Merck E, Mortelmans L, Bormans G, Marchal G, Filez L, Aerts R. Contribution of PET in the diagnosis of recurrent colorectal cancer: comparison with conventional imaging. Eur J Surg Oncol 1995; 21:517-522.
4. Delbeke D, Vitola JV, Sandler MP, al. Staging recurrent metastatic colorectal carcinoma with PET. J Nucl Med 1997;38:1196-1201.
5. Lai DTM, Fulham M, Stephen MS et al. The role of whole-body positron emission tomography with [18F]fluorodeoxyglucose in identifying operable colorectal cancer metastases to the liver. Arch Surg 1996;131:7-3-707.
6. Beets G, Penninck F, Schiepers C et al. Clinical value of whole-body positron emission tomography with [18F]fluorodeoxyglucose in recurrent colorectal cancer. Br J Surg 1994;81.
7. Vitola JV, Delbeke D, Sandler MP, Campbell MG, Powers TA, Wright JK, Chapman WC, Pinson CW. Positron emission tomography to stage suspected metastatic colorectal carcinoma to the liver. Amer J Surg 1996; 171:21-26.

top

Institute for Clinical PET
1820 Jefferson Place, NW
Washington, DC 20036
Telephone: (202) 530-0990 Fax: (202) 833-1613 E-mail:icpmail@aol.com