Role of Video-Assisted Thoracic Surgery in the Treatment of Pulmonary Metastases: Results of a Prospective Trial

Transcription

1 Role of Video-Assisted Thoracic Surgery in the Treatment of Pulmonary Metastases: Results of a Prospective Trial Patricia M. McCormacl MD, Manjit S. Bains, MD, Colin B. Begg, PhD, Michael E. Burt, MD, PhD, Robert J. Downey, MD, David M. Panicek, MD, Valerie W. Rusch, MD, Maureen Zakowski, MD, and Robert J. Ginsberg, MD Departments of Diagnostic Radiology, Pathology, Epidemiology & Biostatistics, and Thoracic Surgery, Memorial Sloan-Kettering Cancer Center, New York New York Background. A retrospective review revealed a 42% error rate between computed tomographic scan reports and thoracotomy findings; therefore, a prospective study was designed to compare the value of computed tomographic scans, video-assisted thoracoscopic exploration, and open thoracotomy in the management of pulmonary metastases. Methods. Eligibility included any patient with only one or two ipsilateral pulmonary metastases identified on computed tomographic scan who was being considered for surgical resection. InitiaLly video-assisted thoracic surgery was performed and all lesions identified were resected. A thoracotomy adequate for complete lung palpation was then carried out and any additional lesions found were removed. Results. Eighteen patients of a planned 50 were treated before closure of the study. Four patients (22%) had no additional lesions found at thoracotomy. The primary sites of tumor were colon (10), breast (3), and one patient each skin (squamous), cervix, kidney, melanoma, and sarcoma. Four patients (22%) did have additional lesions at thoracotomy, which were benign. In the remaining 10 patients (56%) additional malignant lesions were found at thoracotomy after video-assisted thoracoscopic exploration. After 18 patients were entered, analysis of the early results disclosed a 56% failure rate of a computed tomographic scan and video-assisted thoracic surgery to detect all lesions. Being within the 95% confidence interval (32% to 78%), the study was abandoned. Conclusions. We conclude that video-assisted thoracic surgery should be used only as a diagnostic tool in managing lung metastasis. A thoracotomy is required to achieve complete resection, which is the major survival prognosticator for satisfactory long-term results. (Ann Thorac Surg 1996;62:213-7) p ulmonary metastases, when found as the only site of metastatic disease in patients where the primary tumor has been controlled, are often best treated by metastasectomy. The most consistent prognostic factor predicting survival has been complete surgical removal of all metastatic foci. When video-assisted thoracic surgery (VATS) was introduced, it was quickly adapted to metastasectomy. Concern was expressed by us that identification of all metastatic tumors to the lung might not be possible when the VATS technique is used, as manual palpation of the entire collapsed lung has been deemed essential. The use of a probe, or single digit might not su~ce as a substitute for the surgeon's hand [1]. Because the estimation of the number and location of metastases and planned operation is based on computed tomographic (CT) scan images, initially we retrospectively compared the radiologic findings with the surgical findings of 72 patients with primary colon cancer meta- Presented at the Thirty-second Annual Meeting of The Society of Thoracic Surgeons, Orlando, FL, January 29-31, Address reprint requests to Dr McCormack, 1275 York Ave, New York, NY static to lung [1]. Computed tomographic scans underestimated surgical findings in 42% (30 of 72 patients). In those with one or two lesions on the CT scan the "missed lesion rate" was 28%. To validate this retrospective review a prospective study was initiated to determine the accuracy of CT scan and video-assisted surgery as compared to open thoracotomy in identifying metastatic lesions in patients with no more than two presumed pulmonary metastases, believing that these would be the patients eligible for a VATS approach. Material and Methods Any patient who fit the general eligibility criteria for operation for lung metastases, and whose CT scan showed no more than two lesions in one lung was eligible. The selection criteria for metastasectomy are primary tumor under control; no tumor sites outside the lungs; CT scan demonstrates complete resectability; no better proven method of treatment is available; and medical condition allows planned resection. The prospective study was approved by the IRB (93-42) and informed consent was obtained from every patient. Operative decisions made by each surgeon were based 1996 by The Society of Thoracic Surgeons /96/$15.00 Published by Elsevier Science Inc PII S (96)

2 214 McCORMACK ET AL Ann Thorac Surg VATS VS THORACOTOMY IN METASTASECTOMY 1996;62: on scans carried out no more than 3 weeks before the planned operation. The o~icial radiologic report, consultation with radiology staff, and personal review of all imaging material by the surgeon were used for preoperative planning. All CT scans were performed with current generation scanners. Only two scans were obtained helically. Computed tomographic scans more than 3 weeks old or judged unreadable were repeated. The scans were reviewed subsequently by a single member of the Radiology Sta~ (D.M.P.). A data form was used to record the number and locations of all suspected lung nodules seen on lung windows. The degree of confidence in findings was graded on a 5-point scale (0 = normal; 1 = probably normal; 2 = indeterminate; 3 = probably abnormal; 4 = abnormal). Initially thoracoscopy was performed. Any lesion identiffed was resected using standard VATS techniques and instrumentation. Finger palpation and solid probes attempted to identify other nodules. Ultrasonography was not used. Under the same anesthetic a formal thoracotomy was then carried out to allow the surgeon to palpate the entire lung and resect any additional lesions found. All resected specimens were reviewed by one pathologist (M.Z.) and were confirmed to have been completely resected before closure. Results There was no mortality or morbidity in this group of patients and the average postoperative stay was 4.5 days. This was before "same-day admission" policy was in use. Eighteen patients were treated in this fashion before concluding the study. The primary sites were colon in 10 patients, breast in 3, squamous carcinoma of skin, leiomyosarcoma of uterus, cervix, kidney and melanoma were one patient each. In 14 patients, one lesion was noted on CT scan, and in 4 others, two lesions. In 15 patients, all nodules noted on preoperative scans were located at thoracoscopy. In 2 patients with a single lesion on the CT scan, no lesions were identified at the video-assisted procedure. In a third patient only one of two lesions was found and removed at thoracoscopy. All three "missing" lesions were ultimately found and removed at thoracotomy. They were not detectable at VATS because of their deep location and technical problems related to pleural adhesions. Only 22% of patients (4 of 18) had no new lesions found. In 22% (4 of 18) additional benign lesions were found at thoracotomy. In 56% of patients (10 of 18) additional malignant tumors were identified at thoracotomy (Fig 1). In only 1 patient was an additional lesion found at VATS. Of the 14 patients with solitary lesions on CT scan, 7 had additional malignant lesions found (14 nodules). In 3 additional patients five nodules were identified that were benign. In only 8 of 18 patients identified preoperatively as having no more than two ipsilateral lesions was the CT scan correct and confirmed by VATS. Benig 1o New 22.2'~ 22.2% Cancer 55.6% Fig 1. Results of pilot study in 18 patients. In two of the four "two-lesion" patients, only one lesion was found at thoracoscopy. The second malignancy was located at thoracotomy in both patients, one of which had three additional benign lesions discovered. In the other 2 patients both lesions were removed at VATS, but in both additional malignancies were identified and removed at thoracotomy (Table 1). In the single cases with skin, cervix, sarcoma and melanoma primaries, no additional malignant lesions were found at thoracotomy. Twelve of the CT scans were available for retrospective review (Table 2). In 7 of these (58%) our results show a misread, 3 overread, and 4 underread. Comment Pulmonary metastasectomy was first reported in the Scandinavian literature by Divis in 1927 [2] and subse- Table 1. Results of Study Primary Site Method of Discovery Pathology/ Thoracotomy CT scan VATS Thorascopy Cancer Benign Colon Colon Colon Rectal Colon Colon Colon Colon Colon Colon Breast Breast Breast Skin (squamous) Sarcoma (LMS) Cervix Renal Melanoma CT = computed tomography; LMS = leiomyosarcoma; VATS = video-assisted thoracic surgery.

3 Ann Thorac SLrrg McCORMACK ET AL ;62:213-7 VATS VS THORACOTOMY IN METASTASECTOMY Table 2. Retrospective Analysis of Computed Tomographic Scans (12 Scans) Total No. Total No. Primary Site at Operation by Radiologist Breast 3 3 Breast 2 2 Breast 1 1 Colon 8 2" Colon 3 7 b Colon 1 3 b Colon 1 1 Colon 3 2 a Colon 5 2 a Colon 2 I a Skin (squamous) 2 5 b Sarcoma 1 1 Underread; t, Overread. quently in North America by Barney and Churchill in 1939 [3]. However, by 1960, 264 cases of successful resection with long-term survival were reported [4]. In 1971 Martini and colleagues reported 29% 5-year survival rates in osteogenic sarcoma patients where none had survived 3 years before metastasectomy was begun [5, 6]. Multiple nodules were resected and multiple thoracotomies were required to achieve these results. Resection of pulmonary metastases is now the recognized standard of care for selected patients who meet specific criteria. Thoracotomy has been proven to be exceedingly safe with minimal morbidity and mortality [7]. Advanced surgical techniques, first median sternotomy [8] and later the clamshell incision [9] have also made bilateral resections at one operation a safe and efficient approach. This study was designed to compare VATS versus thoracotomy, which is best done with the lateral incision. Therefore, all these patients had thoracotomies. It is our practice to use the median sternotomy and clamshell approach when we are concerned about bilaterality. In sarcoma, with a solitary lesion, we do not explore both lungs. Video-assisted thoracic surgery was introduced as an approach offering a less painful incision, and a shorter hospital stay. It was quickly applied to the treatment of lung metastases [10]. Our initial retrospective study identiffed the potential problems of missing small metastatic lesions, resulting in the institution of this present prospective study. The prognostic significance of several variables in the presentation of these patients has been analyzed to improve the selection of potentially curable patients and avoid operation when it would not help the patient. These variables include length of tumor doubling time and disease-free interval, number of metastases, site of the primary tumor, and completeness of the pulmonary resection. All variables have had inconsistent prognostic significance except one: complete resection of all metastatic lesions [7]. After analyzing the results of the first 18 of a proposed 50-patient study, this protocol was closed to accrual because the estimated probability that cancer will be missed if VATS alone is used is 56% (95% confidence interval, 27% to 75%), and was sufficiently high to warrant termination of the study. Even for patients with only one lesion detected on CT the error rate is estimated to be 50% (95% confidence interval, 23% to 77%). The results of this prospective study are somewhat surprising but do confirm our retrospective analysis-- that in patients with solitary metastases and even more with two metastases--thoracotomy will frequently identify more malignant lesions. Our ability to identify these lesions with a VATS approach alone was less than satisfactory. It is unlikely that the higher accuracy found in the sites other than colorectal, breast, and kidney reflect anything but a sampling error. In this study CT scans were used from a wide variety of referring facilities and represents current clinical practice. It was not practical, due to cost and scheduling, to re-scan each patient with our own helical scanner. We acknowledge that this approach may underestimate the accuracy of CT that can be achieved by consistent, optimized CT technique and a dedicated radiologist comparing old films. It was the only pragmatic solution in our practice climate. In a prospective study of 39 patients, Remy-Jardin and colleagues [11] found 42% more nodules on helical CT scans as compared to conventional scanners. Increased identification of smaller nodules, however, lessens their specificity for being cancers [12]. Retrospective analysis as part of a study favors overreading and our results show a misread in the 7 of 12 patients we analyzed (58%) (Table 2). In conclusion, we firmly believe that metastasectomy has a proven efficacy in treating these stage IV patients with 5-year survival rates in our hands shown to be 80% in testicular primary, 50% in colon or breast, 60% in renal, 40% in osteosarcoma, 25% in soft tissue sarcoma and melanoma primary sites. Patients should be offered operation as an option when they fit the selection criteria. Therefore, we have concluded that the CT scan is not accurate in detecting all pulmonary metastases. The VATS technique as now practiced will fail to detect and remove all pulmonary metastases. Manual palpation of the lung is still required to locate all metastatic foci. Video-assisted thoracic surgery should be used for diagnosis only in metastasectomy. Until newer imaging and localization techniques allow a greater accuracy for the VATS approach, thoracotomy and manual palpation must remain the gold standard in treating even solitary pulmonary metastases. References 1. McCormack PM, Ginsberg KB, Bains MS, et al. Accuracy of lung imaging in metastases with implications for the role of thoracoscopy. Ann Thorac Surg 1993;56: Divis G. Einbertrag zur operativen, behandlung der lungengschuuilste. Acta Chit Scand 1927;62: Barney JD, Churchill ET. Adenocarcinoma of the kidney with metastasis to the lung cured by nephrectomy and lobectomy. J Urol 1939;42:

4 216 McCORMACK ET AL Ann Thorac Surg VATS VS THORACOTOMY IN METASTASECTOMY 1996;62: Gliedman M, Horowitz S, Lewis FJ. Lung resection for metastatic cancer. Surgery 1957;42: Marcove RC, Mike V, Hajek JV, et al. Osteogenic sarcoma under the age of 21: a review of 145 operative cases. J Bone Joint Surg (Am) 1970;51: Martini N, Huvos AG, Mike V, et al. Multiple pulmonary resections in the treatment of osteogenic sarcoma. Ann Thorac Surg 1971;12: McCormack PM, Martini N. A current view of surgical management of pulmonary metastases. In: Economou S, eel Adjuncts to cancer therapy. Philadelphia: Lea & Febiger, 1991: Johnston MR. Median sternotomy for resection of pulmonary metastases. J Thorac Cardiovasc Surg 1983;85: Bains MS, Ginsberg RJ, Jones WG III, et al. The clamshell incision: an improved approach to bilateral pulmonary and mediastinal tumor. Ann Thorac Surg 1994;58: Landreneau RJ, Hazelrigg SR, Ferson PF, et ai. Thoracoscopic resection of 85 pulmonary lesions. Ann Thorac Surg 1992;54: Remy-Jardin M, Remy J, Giraud F, Marquette C-H. Pulmonary nodules: detection with thick-section spiral CT versus conventional CT. Radiology 1993;187: Chang AE, Schaner EG, Conkle DM, Flye MW, Dappman JL, Rosenberg SA. Evaluation of computed tomography in the detection of pulmonary metastases: a prospective study. Cancer 1994;43: DISCUSSION DR MICHAEL J. MACK (Dallas, TX): I congratulate Dr McCormack on conducting this pilot study. I also congratulate her on the occasion of her retirement and acknowledge the valuable contributions she has made to the field of thoracic oncology throughout her career. This specialty is more enlightened because of her contributions. I also appreciate her providing me with a manuscript for review before this meeting. My discussion will focus on two aspects of the presentation: (1) the incisions used and (2) the accuracy of computed tomographic (CT) scanning. As Dr McCormack has alluded to, completeness of resection of all metastatic disease has been said to have positive prognostic significance. Because it is known that up to 45% of patients thought to have unilateral disease by CT scan will in fact have bilateral disease, why then did you not take this concept of complete resection to its logical conclusion and perform a sternotomy or clamshell incision rather than a thora -~ cotomy after video-assisted thoracic surgery (VATS) so that the contralateral lung could also be examined by palpation? The second point is regarding the sensitivity of CT scanning for detecting metastatic lesions. I agree that because surgeons using VATS for resection lose the ability to palpate manually the lung, one is totally dependent on the accuracy of the preoperative CT scan to detect occult nodules. In Dr McCormack's series 78% of patients had additional lesions detected at thoracotomy not imaged on the preoperative CT scan. Unfortunately, 16 of the 18 patients had CT scans performed on old generation scanners rather than on spiral or ultrafast CT scans. In our experience with the newer scans, the problem is not a lack of sensitivity but rather too much sensitivity resulting in detection of nodules as small as 2 mm. A 2-mm nodule on an ultrafast CT scan in a patient with a history of bronchoalveolar carcinoma proved to be a metastatic lesion. However, in a spiral CT scan, in a patient with a history of breast cancer, we saw a 6-mm subpleural nodule. This patient underwent a VATS resection after needle localization. The nodule could not be palpated even in the resected specimen but was found to be a caseous granuloma by frozen section. The results of nodules detected by spiral or ultrafast CT scans and resected by VATS in our institution in 1995 in patients with a recent history of cancer showed that one-third of patients had benign disease, including subpleural lymph nodes, hamartomas, and granulomas. Therefore, I would argue that because of the oversensitivity of the new generation scanners in detecting 2- to 3-mm nodules that cannot be palpated even when the specimen is in the surgeon's hand and because of the higher incidence of benign disease in these small lesions, VATS should have an enhanced rather than a diminished role. My questions for Dr McCormack, therefore, are: (1) Why did you choose a thoracotomy rather than a sternotomy or clamshell incision as a follow-up to VATS? (2) Has there been any evidence of recurrence especially in the contralateral lung in these patients? (3) Do you believe that the study should now be repeated with newer, more sensitive spiral and ultrafast CT scans? DR RODNEY J. LANDRENEAU (Pittsburgh, PA): I also enjoyed Dr McCormack's presentation and congratulate her for her contributions during her career. My comments regarding this presentation are both conceptual and technical. With regard to the concept of complete resection as being a primary prognostic factor with metastasectomy, I believe this is quite a sticky issue and it is vital that we all review the literature carefully when looking at the role of metastasectomy of metastatic lesions to the lung. Many earlier reports mix wedge resection of peripheral solitary metastases with cases of exploratory thoracotomy and aborted resection of bulky disease. Certainly the biology of the metastatic disease in these scenarios is quite different, yet broad statements regarding survival after metastasectomy are frequently made. The accuracy of both VATS and thoracotomy approaches to metastasectomy is limited in that we know that 75% to 80% of all patients will die of their systemic process. But we also know that the use of thoracotomy and aggressive resection commonly results in unnecessary resection of benign lesions and the loss of pulmonary parenchyma. I believe that this is a lot to ask of a patient who is likely to manifest later pulmonary lesions and possible pulmonary compromise from the recurrent malignant process. The indication for VATS wedge resection of pulmonary parenchymal lesions is quite specific. It should be directed toward small peripheral lesions that can be encompassed by complete resection of the lesion. There should be no endobr0nchial extension of the disease, and finally, we must always remember that conversion of thoracotomy is necessary when the lesions identified by CT scan cannot be found. I end by saying that metastasectomy done by thoracotomy or VATS is in almost all circumstances diagnostic only. We must remember that our first duty to our patients suffering from the spread of their primary malignancy is to do no harm. In the case of metastasectomy, we must be careful to avoid chasing the tail of their systemic disease. DR RALPH J. LEWIS (New Brunswick, NJ): I too congratulate Dr McCormack on her retirement and, actually, a very fine and interesting presentation that was almost convincing. The literature is very confusing, contradictory, and, at best, anecdotal for curative, surgical resection of multiple, epithelial, pulmonary metastases, and I emphasize epithelial metastases.

5 Ann Thorac Surg McCORMACK ET AL ;62.'213-7 VATS VS THORACOTOMY IN METASTASECTOMY We have never had such enthusiasm for this procedure because of consistent and even predictable poor outcomes, but for those surgeons persuaded that this can be a curative operation, VATS could still be beneficial. It can be used bilaterally at the same sitting, and it even mandates a no-touch, oncologic technique, avoiding the implied release of showers of malignant cells into the bloodstream from extensive palpation. I would just like to say a word or two about the CT scan. It can diagnose very, very small parenchymal lesions, 2 to 3 mm in size. Maybe some lesions so small that they cannot be delineated by palpation alone. As we know, many metastases occur in the periphery of the lung and even on the visceral pleural surface. My radiologists have informed me that these small surface lesions are not detected accurately by the CT scan, and they can be easily missed. Therefore the CT scan, by itself, can only be expected to visualize a portion of the true number of metastases present. This might explain the disparity in trying to count lesions preoperatively, and the number that is actually removed at operation. If we use the magnification of the VATS endoscope, possibly, some very small surface lesions, that could not be seen by the naked eye alone might be detected. Therefore by combining these two technologies, magnification of VATS with the CT scan, probably, we could get a more realistic evaluation. I would like to ask two or three questions. First, are you really repudiating the no-touch, oncologic technique for neoplasms and instead advocating an extensive, compressive, palpation method that could cause the release of malignant cells into the bloodstream? And second, we have all been told that wedge resection is really not an oncologic operation, at least for that very small peripheral, primary, T1 NO lesion, and yet, you seem to be sanctioning wedge resection for bilateral, multiple, large lesions of varying sizes. Also, I am confused because you are performing only unilateral thoracotomies. Why are you accepting the CT scan as being accurate and reliable for metastasis to the contralateral lung? In the literature, some patients have had two, three, and four separate thoracotomies for newly appearing metastases. For these particular patients, would not VATS be more appropriate, kinder, gentler and even more humane? Finally, once again, I congratulate you on your many contributions and wish you a long and very happy retirement. DR McCORMACK: I thank all of the discussers for their kind remarks, especially Michael Mack. And as everything seems to be focused around the same points, let me just say I think Ralph Lewis summed it up probably best: Those who are enthusiastic for this approach to treat metastatic cancer should do so. Those who are not enthusiastic about this approach for metastatic cancer should not do this type of operation. As for incisions, we selected a thoracotomy because it was part of the methodology of this study. Practically speaking, most of us now are much more liberal with the use of a median sternotomy or a clamshell incision to get all of the lesions out bilaterally at the first operation. Recurrences have been found in either lung, the operated lung or the opposite lung. I cannot report on this particular study, it is too short, but our experience over 40 years has been that recurrence is unpredictable. Surface lesions are those that would be picked up, Dr Lewis, by the video, the VATS technique, very nicely. What we have found in this study is that most of the additional lesions are deeply situated in the lung and are not on the surface. The proper selection of patients is used in the criteria selection for each patient. This would weed out those patients who would not profit by another surgical procedure to extend their life. In most of our metastasectomy selections, we have discovered that nothing else works for these patients. Just as an example, in colon metastases, nobody survives beyond 24 months with chemotherapy, and we have a 5-year survival of 50%. This is indeed a technique worth doing, and I think for the sake of the patients it should be offered to them. Cancers that are metastatic to the lung we found grow differently from primary lung cancers; therefore, a wedge resection is an adequate surgical technique. We have not found that by carefully palpating a collapsed lung we have spread cancer cells. Bound volumes available to subscribers Bound volumes of the 1995 issues of The Annals of Thoracic Surgery are available only to subscribers from the Publisher. The cost is $99.00 (outside US add $25.00 for postage) for volumes 59 and 60. Each bound volume contains a subject and author index, and all advertising is removed. The binding is durable buckram with the name of the journal volume number, and year stamped on the spine. Payment must accompany all orders. Contact Elsevier Science Inc, 655 Avenue of the Americas, New York, NY 10010; or telephone (212) (facsimile: (212) ).