| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 1997;64:776-777
© 1997 The Society of Thoracic Surgeons
DR FRANCIS G. DUHAYLONGSOD (Durham, NC): I thank Dr Block for an excellent presentation.
Over 2
years, the Washington University group has evaluated 58 patients with biopsy-proved esophageal cancer. All patients underwent computed tomographic (CT) and positron emission tomographic (PET) imaging. On the basis of the preoperative evaluation, patients were classified into one of three groups: inoperable, unresectable because of the presence of distant metastatic disease, or resectable. Because the accuracy of whole-body PET imaging to detect distant metastases is well established in the Duke experience (more than 700 lung cancer patients to date), I will limit my comments to the resectable group only.
Thirty-five patients were considered potentially resectable and underwent operation. Twenty-nine regional lymph node metastases were found in 21 patients. The results of PET imaging were 13 true-positive, 11 true-negative, 3 false-positive, and 16 false-negative. The results of CT included 6 true-positive, 11 true-negative, 3 false-positive, and 23 false-negative. Thus, the sensitivity of 2-[18F]fluoro-2-deoxy-d-glucose (FDG)–PET to detect regional lymph node metastases was 45% versus 21% for CT. The specificity was 79% for both modalities. The accuracy was only 56% for PET versus 40% for CT. Consequently, from a purely statistical standpoint, PET is superior to CT. From a clinical standpoint, however, the efficacy of FDG-PET to detect regional lymph node metastases is not appreciably better than tossing a coin.
In a recent publication from our institution, detection of mediastinal nodal metastases in patients with lung cancer using FDG–PET demonstrated a sensitivity of 90%. How can we reconcile these disparate values? I suggest three possibilities: (1) difficulties with subjective interpretation of FDG-PET imaging; (2) difficulties with correlative lymph node mapping; and (3) difficulties in resolving increased FDG uptake in lesions less than 1 cm in diameter.
All PET images in this study were evaluated by experienced nuclear medicine physicians; however, no objective criteria were given to define a positive, negative, or equivocal PET study. Hence, extrapolation of these findings to other institutions is virtually impossible. My first question is, Is there any value to semiquantitative tumor uptake indices to improve detection of nodal metastases in patients with esophageal cancer?
Twenty-two of the 35 patients who underwent resection had a transhiatal esophagectomy. The completeness of regional lymphadenectomy is less with the transhiatal approach than with systematic en bloc esophagectomy. Indeed, even with meticulous surgical dissection and labeling it is very difficult to assure that individual biopsy specimens correspond precisely to the radiographically mapped lymph nodes. My second question is, Could distortion of the anatomic relationships and the limited node dissection adversely affect the accuracy of PET imaging?
In our experience, the accuracy of FDG–PET is moderately impaired in lesions less than 1 cm in size. My final question is, What was the size range of the regional lymph nodes in the surgical specimens? How did the sensitivities compare between lesions less than or greater than 1 cm in diameter?
I applaud the Washington University group for pioneering the use of FDG-PET for the evaluation of esophageal cancer, but some caution is warranted. At present, the poor sensitivity of PET will subject some patients with regional lymph node disease to a major operation alone without the putative benefits of neoadjuvant therapy. Nonetheless, this should not detract from the importance of this paper as a springboard to further study. Doctor Block, I look forward to future presentations by your group.
DR THOMAS M. EGAN (Chapel Hill, NC): I congratulate the authors on a very well presented study. I have a specific question about defining patients as having unresectable disease on the basis of PET scanning and whether that was done because of the presence of celiac nodes.
In the paper we presented yesterday, using neoadjuvant therapy, we had a number of patients with enlarged celiac nodes who were included in the radiation port as part of the preoperative irradiation and chemotherapy protocol. At the time of resection, these nodes were resected en bloc with the proximal portion of the stomach. Did you exclude from resection patients who had evidence of celiac node positivity on PET scan?
DR BLOCK: Thank you very much, Dr Duhaylongsod and Dr Egan. I will address Dr Egan's comments first. There was 1 patient who had positive celiac lymph nodes by both PET and CT and was therefore considered to have unresectable disease. No patient had an isolated finding of celiac node positivity by PET as the sole determinant of unresectability.
I appreciate the very insightful comments of Dr Duhaylongsod. They get to an important consideration in the use of PET for the staging of cancer—that of sensitivity and specificity. I will address your comments with this fundamental question in mind: What is it that we are trying to achieve with our preoperative staging evaluation?
This presentation was intended to illustrate that in a substantial proportion of patients evaluated, PET could identify unresectable disease missed by conventional staging investigations and thereby prevent a surgical intervention that would be of questionable benefit to the patient. The study was not designed to evaluate the sensitivity, specificity, and accuracy of PET. Therefore, I approach your three questions with that in mind.
First, is semiquantitative analysis of value? Our nuclear medicine colleagues, who have extensive experience with interpreting PET images for a variety of malignancies, have chosen to avoid the use of semiquantitative analysis in evaluating patients with cancer of the esophagus. Their concern relates to intense activity by the primary tumor in the mediastinum interfering with quantitative analysis of mediastinal lymph nodes. Further, the primary goal was to identify sites of distant metastases indicative of unresectable disease, not to identify involvement of local lymph nodes. Doctor Luketich yesterday presented similar overall results using a quantitative approach to interpretation of PET images. Certainly that approach eliminates physician experience as a variable, and perhaps it facilitates broader application of PET in centers with less experience. I hope, however, that at centers that invest in PET, nuclear medicine physicians will rapidly develop the expertise necessary to allow their interpretation of the images to enhance the scan's sensitivity.
The second question relates to the impact the technique of esophagectomy has on accurate analysis of mediastinal lymph nodes. Certainly a transhiatal esophagectomy presents greater difficulty for complete removal of nodes and for preservation of anatomic relationships, thus making correlation of preoperative conclusions with pathologic findings less secure. Again, however, this affects the calculated sensitivity and specificity, end points unrelated to the purpose of the study. The importance of this issue depends on the answer to our original question about the role of PET as a staging modality and the clinician's approach toward treating patients with this disease. Our results suggest that PET is not yet able to distinguish reliably between the patient with involved adjacent lymph nodes and the patient with an N0 lesion. For patients in whom the disease appears resectable, consideration of neoadjuvant therapy on the basis of involvement of adjacent nodes must be addressed with additional information. Our conclusion, and we believe the real value of PET, relates to identifying those patients for whom the presence of unresectable disease renders neoadjuvant therapy and attempted resection moot considerations. We therefore chose not to include the data on lymph nodes in the resected specimens in today's presentation.
Third, you asked about the size of lymph nodes and whether or not that had a bearing on their detectability by PET. We did not review our data to evaluate the size of lymph nodes and whether or not PET was able to detect them. That is an endeavor fraught with difficulties because to be thorough, one must systematically approach every identifiable, or potentially identifiable, lymph node, ensure a biopsy is done, and compare the findings with that specific site on the preoperative studies—a task of questionable reliability at best. The study design did not include this type of analysis, and therefore we cannot say whether large lymph nodes were detected more often than small ones. I certainly suspect that would be the case.
The resolution of PET is said to be approximately 10 mm at present. Future-generation scanners will have better resolution, but it seems intuitive that small lymph nodes, analogous to the two very small primary tumors in 2 patients, are liable to escape detection.
Related Article
Ann. Thorac. Surg. 1997 64: 770-776.
This article has been cited by other articles:
|
|
 |
|
  M N Maisey Overview of clinical PET Br. J. Radiol., November 1, 2002; 75(90009): S1 - 5. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
