- Email: [email protected]
Primary Cancer of the Lung
Primary Cancer of the Lung A 42-Year Experience Richard H. Overholt, M.D., Wilford B. Neptune, M.D., a n d Mian M. Ashraf, M.D. ABSTRACT Between April, 1932, and July, 1974,3,808 patients with primary lung cancer were studied and 1,848 underwent resection. Among untreated patients, 95% were dead within a year. Unresected cancer of the lung is so lethal that efforts to streamline surgical management should not be neglected. In good-risk patients with isolated lesions the approach can be direct. If surgical excision is indicated, regardless of a positive or negative sputum cytology, bronchoscopic biopsy, or brush biopsy, such investigations become superfluous. Needle biopsy is also inconclusive and in addition is hazardous. Preoperative investigation should focus on cardiopulmonary reserve more than on ways to obtain tissue for verification. With the passage of time, the extent of resection has become more conservative. The value of palliative resection is now better appreciated in terms of quality of life, its prolongation, and, for some, a possibility for cure.
P
rimary cancer of the lung was rare and always fatal before the early thirties. It is now common a n d selectively curable. A dismal picture, however, has been presented by most publications. They headline low overall cure rates, and even reports on results of mass screening programs are gloomy. During the past decade the majority of papers have focused on palliative programs of irradiation or chemotherapy. They advocate exhaustive attempts to obtain tissue or cells from sputum o r by extraction through endoscopic or direct needle aspiration. Many current articles stress the importance of avoiding operation. Our experience suggests that such intervention is necessary not only to establish the existence of the cancer in its early stages, but also to determine its extent in the advanced case.
Clinical Material and Methods Over more than four decades, 3,938 patients suspected of having primary malignancy of the lung have been observed and 3,808 have had a tissue diagnosis. The policy of management has largely been influenced by the senior author, and shifts in staff have been minimal: a total of 8 surgeons have been involved. Only histologically verified cases are the subject of this report. All bronchial adenomas, 175 in number, are excluded. T h e reference background is unique. In the first From the Overholt Thoracic Clinic, New England Baptist and New England Deaconess Hospitals, Boston, Mass. Presented at the Eleventh Annual Meeting of The Society of Thoracic Surgeons, Montreal, Que., Canada, Jan. 20-22, 1975. Address reprint requests to Dr. Overholt, Overholt Thoracic Clinic, 135 Francis St., Boston, Mass. 02215.
VOL. 20, NO. 5, NOVEMBER, 1975
511
OVERHOLT, NEPTUNE, AND ASHRAF three decades, many patients were referred from older populations of tuberculosis sanatoria. Twenty-four percent of those who underwent resection had diabetes or an associated serious disease. All patients with a verified diagnosis (many of whom were terminal) seen in consultation with other surgeons are included. Sixty-nine patients had a second resection, and 1 of these had a third. There were 37 ipsilateral and 32 bilateral operations. In 4 instances the second resection was either a pneumonectomy in a patient who had previously undergone resection of a portion of the contralateral lung or segmental resection of the remaining lung in a patient previously treated by pneumonectomy. Therefore this series of patients, like any other, has built-in variables as to selectivity. The first patient, treated in April of 1932 [2], was a 59-year-old man with severe chest pain. He had a necrotic epidermoid lesion the size of a grapefruit in the right lower lobe, and a palliative lobectomy was done. The lesion had invaded the pericardium. Convalescencewas uneventful; the hospital stay was 17 days. He received postoperative irradiation and spent a comfortable 10 months, then died suddenly while eating breakfast. Since then, of all the patients seen, 1,182 were rejected for operation, 57% were explored, and 48% of the total had resection. This summary focuses largely on the last group, numbering 1,848 patients. In the entire series the ages covered a spread from 15 to 93 years (average, 59) with the majority of patients between 45 and 70 years. The youngest resected patient was 15 and the oldest, 86 years. Clinical staging in retrospect according to the American Joint Committee recommendations of September, 1973 [ 11is impossible. Patients have simply been divided into two groups: those whose cancer was strictly localized to the primary site and those with evidence of spread either to a lymph node or by direct extrapulmonary extension. Of the resections performed, 1,162 were considered palliative and 686 curative. The resected tumors were adenocarcinoma in 17%of the patients, epidermoid in 4476,undifferentiated in 28%, .small cell in 5%, and other cell types in 6%. With time, the extent of resection has become more limited. Two factors are responsible: lesion size and acceptance of more patients with a low pulmonary reserve (Fig. 1).The Table shows operative mortality in the early and late periods of our experience according to the extent of resection as well as the objective of the procedure, i.e., for cure or palliation or only for exploration. During the first FIG. 1 . The resection experience is divided f o r convenience into four periods of ten years each with two extra years added to thefirst period. The change in the extent of resection performed is illustrated by decades. ( s e g = segmentectomy; lob = lobectomy; pn = pneumonectomy.)
%
25 I
512
T H E ANNALS OF THORACIC SURGERY
Primary Cancer Ofthe Lung OPERATIVE MORTALITY IN 1,848 RESECTIONS FOR LUNG CANCER
Mortality (%)
Procedure
Early
(1932- 1954)
Late (1955- 1974)
EXTENT OF RESECTION
Segmentectomy Lobectomy
Pneumonectomy
25 11
17
3
5 14
GOAL OF OPERATION ~~
Exploration only Curative resection Palliative resection
4 11 14
3 5 12
decade of this experience, operations were performed without the benefit of a blood bank, antibiotics, steroids, ventilatory assist devices, electronics, or a recovery room or intensive care unit. Blood from a bank was available in 1941, then penicillin in 1944. The other refinements came during the last half of this experience. All but 29 of the patients who underwent resection have been followed to the time of death or to survival for more than five years. Data on five-year survival when follow-up was possible revealed that 95% of the 1,960 untreated patients died in the first year after diagnosis (average survival, 4.2 months for patients rejected for operation and 6.1 months for those who were only explored). Results in treated patients are recorded according to the histology (Fig. 2), extent of resection (Fig. 3A), palliative or curative resection (Fig. 3B), and combinations of therapy (Fig. 4). Either alone or in combination, 747 patients received irradiation and 234 had chemotherapy.
Comment Out of all this experience, many lessons have been learned, but comment will be limited to three important points: the circumstances of discovery, simplification of the approach to the localized lesion, and development of a more direct approach to disseminated disease. In our series, only 3 of the cured patients were apprehended at a time when the chest films were normal. These patients all had symptoms that led to sputum investigation or bronchoscopy with biopsy, lavage, or brush. All the asymptomatic patients were discovered by radiological survey. Massive population screening for early cancer by cytological examination of bronchial secretions does not appear to be practical. In the past, the medical profession -including many radiologists - has not taken small, even minute areas of abnormal uncalcified density seriously enough VOL. 20, NO. 5, NOVEMBER, 1975
513
OVERHOLT, NEPTUNE, AND ASHRAF % 100
75 5 0
2 5.
5Yr
FIG. 2. Five-year survival according to histological type in patients who underwent resection shows that the outcome with adenocarcinoma (adeno) is 33 slightly more favorable than that with efndermoid 26 carcinoma (epid). The one-, two-, and three-year 22 survival exceeded by three times that of the untreated group. There were no long-term survivors among the 0 small or oat cell group. (undiff = undifferentiated carcinoma.) Sr
in terms of their cancer potential. T h e interval between the initial discovery film or onset of symptoms and the surgical investigation has been too long. This interval in the first decade of our experience was 18 months; in the second, 10 months; in the third, 9 months; and in the last, 5 months. Three paradoxes exist which indicate the need for a more direct course of management: (1) the fewer the symptoms and abnormal signs, the greater the likelihood that the substance producing a roentgenographic shadow is a cancer; (2) curability is inversely proportional to the number of symptoms and signs; and (3) the easier it is to obtain specimens for verification, the worse the prognosis. Therefore, if a person harbors a cancer within the lung, the most fortunate circumstance for him would be to have the shadow of the cancer discovered by chest roentgenogram at a time when he is apparently in robust health and free from symptoms or abnormal physical signs, with laboratory studies all within normal limits, the lesion not visible bronchoscopically, and the cytology negative regardless of how the specimen was obtained. The highest cure rate is attained in this set of circumstances. For isolated lesions in good-risk patients the approach can be direct. If surgical excision is indicated, regardless of a positive or negative sputum, bronchoscopy, biopsy, or brush, such investigations become superfluous. Needle aspiration of a suspicious area in the lung is also inconclusive and in addition is
7 5.
66
50.
40
2512
FIG. ?. Survival in 1,582 patients based on (A) extent of resection and (B) palliative or curative resection. The less extensive the resection, the better the survival. Note that life was polonged beyond one year for 40% of the resected patients (as compared with 5 %f o r those untreated). One hundred twenty-nine cases were discovered by survey; the survival rate was highest in this group. (seg = .regmentectomy; lob = lobectomy; pn = pneumonectomy.)
514
T H E ANNALS OF THORACIC SURGERY
Primary Cancer Ofthe Lung
15
FIG. 4 . ( A ) Palliative treatment consisted of a combination of chemotherapy and irradiation in 7 1 patients. In 2 7 of these it was possible to do a resection also. The triple combination produced better results in terms of prolongation of lfe. ( B ) Supervoltage irradiation was given to 478 patients, 200 of whom also had a resection. (C) One hundred j$y-seven patients were treated @ chemotherapy, and of these 77 also had a resection. Obviously, those who could be resected had less advanced disease, and results were expected to be better. However, in those patients treated only by chemotherapy, the survival curve ZS almost identical to that for the untreated group.
hazardous. Preoperative investigation should focus on cardiopulmonary reserve more than on ways to obtain tissue for verification. In patients with small, localized lesions who are asymptomatic or relatively so, little preparation is indicated. In the first decade of our experience, the relationship of smoking to chemical bronchitis and emphysema was less well appreciated than it is today. The principal etiological factor that initiates the cancer also diffusely damages the entire lung. Fortunately, chemical bronchitis is reversible. Cessation of smoke inhalation is the only preparation necessary for most patients in this category. The issue in disseminated cancer should also be settled directly and positively. Isotopic screening of skeleton or liver may give presumptive evidence of dissemination. Bone marrow aspiration may be indicated. The routine in many clinics is a series of studies performed several days apart. The investigation proceeds from one to the next test until positive tissue or cells are obtained. Bronchoscopy usually comes first, then cervical node biopsy, then percutaneous needle aspiration of the lesion, and finally mediastinoscopy or direct mediastinal exploration. Multiple iqyestigations can often be scheduled for the same day with a pathologist present for frozen section. In recent years, the first approach has often been directly to the liver or retroperitoneal nodes by laparotomy. If involvement is found below the diaphragm, either operation or irradiation of the lung would be futile gestures as far as hope for cure is concerned. At this point in time, surgical removal continues to provide the best hope for prolongation of life and possible cure. It is therefore imperative to focus on factors
VOL. 20, NO. 5 , NOVEMBER, 1975
515
OVERHOLT, NEPTUNE, AND ASHRAF which are basic to the surgical approach. The surgeon is responsible for determining the identity as well as extent of the lesion. Both should be accomplished in the most direct and absolute way possible. As one perfects, one simplifies. For the potential victim with a localized lesion and an adequate pulmonary reserve, histological verification and total biopsy to determine the extent of the cancer settle both issues and affect a cure simultaneously. Multitudinous preoperative investigations then become superfluous. The surgeon must also settle the issue when the lesion has extended beyond the primary site. Again, the most direct approach for verification and determination of extent should be taken. Hopelessness should be established early so that the patient can be spared suffering, expense, and useless days in a hospital if the terminal period is approaching. This experience has taught us to be more conservative in the extent of resection for small localized lesions. The value of palliative resection is now better appreciated in terms of prolongation of life, its quality, and, for some, the actuality of cure.
References 1. American Joint Committee for Cancer Staging and End Results Reporting. A Clinical Staging System for Carcinoma of the Lung, Chicago, 1973. 2. Overholt, R. H. Primary carcinoma of the lung: Surgical extirpation. A m ] Surg 22:181, 1933.
Discussion DR.J. GORDON SCANNELL (Boston, Mass.): Even though both Dr. Shields and Dr. Overholt made their manuscripts available to me in advance, it is clearly a challenge to discuss these two papers on cancer of the lung when one realizes that some 6,000 patients with a “surgical” type of disease are involved. Since one study is collaborative and the second is a personal experience, they have to be dealt with separately. The conclusions reported in the first paper by Shields and his group are an interesting by-product of a study designed for another purpose, namely, the evaluation of adjuvant chemotherapy. There is therefore an important element of preselection in it: only those who could be resected are considered. Under the circumstances, it is really not too surprising that cell type in itself is not a very meaningful determinant, particularly if one excludes oat cell carcinoma. I assume that bronchial adenoma has been omitted - or at least not tucked away into any of the groups -for though it may indeed be a malignant disease, it is not cancer of the lung in the ordinary clinical sense. In the same way, I would like to be sure that bronchiolar and alveolar cell carcinoma are not included. These remarks are not made in the spirit of captious criticism. I think the authors are very much to be commended for reporting their results in terms of staging of the disease. I do find it very interesting that they can distinguish so clearly between hilar and mediastinal involvement. Is there a mythical boundary that perhaps some of us have difficulty seeing? Or is it, in effect, a quantitative description of the degree of lymphatic involvement? I confess to a certain disappointment that postoperative irradiation in the presence of positive nodes was not included in the study. This, to me, still remains an important unanswered question in the surgical management of cancer of the lung, although I recognize that very real steps are being made by the Michigan group in this field. Personally, I have the clinical conviction (or prejudice) that postoperative radiation therapy in the presence of positive nodes, whether squamous or adenocarcinoma, has a real place in the management of the patient, and probably also of the cancer. Now that mediastinoscopy permits more accurate preoperative staging, the relative merit of postoperative versus preoperative irradiation, which still is a somewhat unanswered question, becomes important. 516
THE ANNALS OF THORACIC SURGERY
Primary Cancer of the Lung You really have to change gears to go from a collaborative study to the experience of a surgeon who has made cancer of the lung one of the personal challenges of his long, distinguished, and active career. It is all too easy - as I think Dr. Overholt implies -to depersonalize cancer of the lung, and I sense some impatience with too elaborate efforts to reconnoiter the field before launching an attack. One cannot help agreeing with a great deal of this, but I do feel that if we are going to make progress in defining the place of surgery, we will have to do more than divide operations into the categories of those done for “cure” and those for “palliation.” Both terms are in the eye of the beholder, in this case the operating surgeon. What is “palliative” is a matter of judgment, as shown by the fact that some 13% were actual survivors. But to evaluate the results of operation only in terms of cure is probably too harsh a judgment - again, as Dr. Overholt has implied. I would like to think (and I admit to a certain amount of chauvinism in this) that the reason limited anatomical resection -that is, lobectomy - became acceptable is attributable in no small measure to the ideas of my former mentor and Dr. Overholt’s contemporary, Dr. Edward Churchill. Certainly, both have taught us to carry the battle to the enemy, and if I may pontificate for the moment, both have accepted St. Paul’s injunction: “If the trumpet give forth an uncertain sound, who will prepare for the battle?” DR. DONALD L. PAULSON (Dallas, Tex.): This is really a remarkable morning, and I am privileged to follow Dr. Overholt’s excellent classic paper covering more than forty years’ experience by discussing Dr. Shields’ paper. When we examine separately the influence of cell type on prognosis, we find a strong relationship, particularly in the case of well-differentiated epidermoid carcinoma in contrast to very undifferentiated small cell carcinoma. However, taking all the variables together, histological type is of secondary importance to the stage and extent of disease. The relationship of histological type and mediastinal lymph node metastasis is noteworthy, because in both anaplastic carcinoma and some adenocarcinomas there is a much higher incidence of detected involvement than in the epidermoid cell type. Once the nodes are involved, however, the prognosis is poor, though variable in its course. Much of the confusion regarding the meaning of mediastinal node involvement arises from a lack of precise delineation of the location and extent of nodes involved. The location is characterized as being ipsilateral or contralateral, high or low peritracheal, and tracheobronchial. Those patients with contralateral or high peritracheal spread are usually inoperable, and their life expectancy is less than two years. Only those with involvement of the tracheobronchial nodes may be considered suitable for operation. In our series of 251 resections in patients with mediastinal node involvement, only 15 survived five years. All of these were patients with ipsilateral tracheobronchial nodes involved intranodally. The extent of node involvement, i.e., intranodal or perinodal, is very relevant, as has been shown by Dr. Bjork. Bergh and Schersten found that the operative mortality for patients with perinodal metastases is about five times greater and the survival one-tenth as good as in those with intranodal metastases. Unfortunately, perinodal involvement predominates in all histological types, but ipsilateral intranodal growth is found in about 12% of all cases. The operative mortality corresponds well with the established five-year survival rate of less than 10% after resection in patients with mediastinal node metastases. We do not think patients with mediastinal node involvement are automatically inoperable, but that there are some patients better suited for operation, the favorable lesions being those in the presence of mediastinal metastases that are ipsilateral, low peritracheal, or tracheobronchial in location and intranodal in extent of involvement. DR.MARVINM. KIRSH(Ann Arbor, Mich.): Since 1955, superior mediastinal, subcarinal, peritracheal, and paraesophageal mediastinal node dissection has been carried out following pulmonary resection at the University of Michigan Medical Center. If these nodes were positive, the patients received postoperative radiation therapy. The overall five-year survival for the group undergoing curative resection has been around 30%. In contrast to Dr. Shields’study, our results suggest that histological cell type is an important factor in determining the prognosis in patients with bronchogenic carcinoma, especially if there is evidence of lymph node metastasis. As a group, patients with squamous cell carcinoma of the lung had a greater five-year survival than patients with VOL. 20, NO. 5, NOVEMBER, 1975
517
OVERHOLT, NEPTUNE, AND ASHRAF adenocarcinoma of the lung. Even in the absence of lymph node metastasis, patients with squamous cell carcinoma tended to survive longer than those with adenocarcinoma. The difference in survival between the two groups was more striking when there were hilar or mediastinal node metastases. The five-year survival rate in patients with squamous cell carcinoma and hilar metastases treated by pulmonary resection alone was 45%, and it was 29.5% in patients with mediastinal metastases treated by pulmonary resection, mediastinal dissection, and postoperative mediastinal irradiation. In contrast are the poor results in patients with adenocarcinoma and lymph node metastases. None of the patients who had adenocarcinoma with hilar metastases lived five years, and only 5.9% of those with adenocarcinoma of the lung and mediastinal node metastases survived five years with operation, node dissection, and irradiation. In addition, this study also suggests that in patients with squamous cell carcinoma of the lung, it is the presence and not necessarily the extent of metastases that determines the prognosis, provided that patients with squamous cell carcinoma of the lung and mediastinal metastases receive radiation therapy postoperatively. Because of the poor results we have achieved with adenocarcinoma of the lung and mediastinal metastases over the past ten years, we are currently reevaluating our approach for these individuals. DR. JAMES J. YASHAR(Providence, R.I.): During the six-year period from 1963 to 1969, 148 patients underwent thoracotomy for bronchogenic carcinoma. We had a resectability rate of 84%; this is in contrast to the 48% reported by Dr. Overholt. Several factors influenced the survival rate. First, the tumor size: with tumors less than 2 cm in diameter the five-year survival was 61%; for those between 3 and 4 cm it was 36%; and between 5 and 7 cm, survival was 27%. We had no survivors among the patients who had a tumor greater than 7 cm. Our overall five-year survival was 34%. Second, lymph node involvement exerts an unfavorable influence on lung cancer. With no node involvement the five-year survival was 44%; with interlobar involvement, 33%; and with hilar involvement, 25%. We had no five-year survivors among patients who had mediastinal node involvement. This is statistically significant. The third factor is histological type. Our data indicate that patients with epidermoid carcinoma stand a better chance of having a resectable lesion. The survival rate for squamous cell carcinoma was 35%; for undifferentiated, 30%; for adenocarcinoma, 40%; and for bronchiolar carcinoma, 50%. These differences are not statistically significant. We grouped patients with oat cell carcinoma in a different category since these tumors behave differently biologically; we had no survivors among the patients who had an oat cell carcinoma. I would like to ask Dr. Shields and Dr. Overholt whether, based on their experience, they think there is any relationship between tumor size and five-year survival. DR. R. A. RASMUSSEN (Grand Rapids, Mich.): Dr. Overholt has presented a 42-year experience with 3,900 cases of lung cancer, emphasizing a high incidence in the fifth and sixth decades with an increasing percentage of women and a continuing need for early surgical consideration in all cases. He obtained a 48% five-year survival in favorable cases. This is an encouraging report from an experienced surgeon. Too much bad news, gloom, and doom are expressed by inexperienced workers, many of whom seem obsessed with a positive tissue diagnosis even though their methods are costly, time consuming, inadequate, and often as dangerous as diagnostic thoracotomy, which also offers the opportunity to perform an appropriate definitive resection. We must reorient the profession. There is good news about lung cancer. We know the cause -that it is seldom seen in nonsmokers and is therefore largely preventable. Periodic chest roentgenograms in smokers offer early diagnosis and detection. Once the tumor is detected and proper surgical resection accomplished, there is hope of a cure with good survival. We must emphasize both the use of diagnostic thoracotomy instead of all the lesser diagnostic techniques and avoidance of dangerous decisions on operability by generalists, internists, and radiologists. We will then improve our accuracy. Presently, surgical therapy offers the only real hope for victims of lung cancer. We find nearly 50% five-year survival in the favorable resected patient, and often there is surprising survival even in unfavorable 518
THE ANNALS OF THORACIC SURGERY
Primary Cancer of the Lung cases. Careful follow-up is necessary to ensure that death is due to persisting lung cancer and not to some other cause such as heart disease or the postpneumonectomy state. The picture is reasonably bright. In 1963 we reported a Western Michigan experience with 813 patients, emphasizing that lesser resections were as curative as pneumonectomy. We now have approximately 2,500 cases and maintain this position. I would affirm Dr. Overholt’s report and hope that we heed his suggestions. DR. SHIELDS:First, bronchial adenomas were omitted from the studies throughout. Alveolar cell tumors were excluded in the initial studiesand then later on were placed in the category of adenocarcinoma, since that is exactly what they are. One of the things that impresses me in all discussion on lung carcinoma is that great differences in long-term survival are reported - from as high as 48% to as low as 8% despite the fact that we are talking about the same disease. This can stem from only three things: selectivity in patient population, selectivity and methodology of pathological examination, and very small numbers of patients being used to determine long-term survival. Many of the results with small numbers are not significant. One of the spin-offs of the various collaborative studies is that a large number of patients are selected without personal prejudice and an evaluation of treatment is more readily possible. Concerning a comment by Dr. Scannell, when one is doing a randomized study to evaluate the use of adjuvant cancer chemotherapy, one does not have the option to use postoperative irradiation even if the mediastinal nodes are positive. However, I agree that such a study should be done, and one of the surgical protocols of the Working Party for the Therapy of Lung Cancer has been set u p to evaluate irradiation when the mediastinal nodes are diseased. Only from such a prospective randomized study will the true value of postoperative radiation therapy be found. In the past it has been used on a selectivebasis, so that its true value is undetermined; but certainly this adjuvant must be looked into. I believe it is extremely difficult to talk about tumor size except with peripheral lesions, since in central lesions distal alterations in the parenchyma due to inflammatory changes behind the tumor itself make it difficult to evaluate the true size. In the peripheral pulmonary nodule study that was conducted initially by Dr. John Steele, size was noted to be important: patients whose lesions were less than 3.5 cm had a much better prognosis than those with larger lesions. Also in this study, a small number of patients with oat cell tumors underwent surgical resection. (We believe peripheral oat cell tumors that have no nodal involvement should be resected.) T h e five-year survival for peripheral oat cell lesions was essentially the same as for adenocarcinoma, squamous cell carcinoma, and the undifferentiated large cell types. My last comments concern terminology. I don’t think the word palliation should be used to describe a resection (even though mediastinal nodes are involved) if all known tumor is removed. I believe it should be classified instead as curative resection. When tumor is left behind, the resection should be called incomplete; use of the term palliative in this situation is inappropriate. DR. OVERHOLT:Dr. Scannell raised a question about the exclusion of adenomas. These were omitted; there were approximately 90 of them. Also, I resected the right lung of a patient in November, 1933, and our pathologist called the tumor an adenocarcinoma. Later the diagnosis was changed to bronchial adenoma, and this case is not included. Dr. Scannell also questioned bronchiolar and alveolar types. We grouped those under adenocarcinoma. Dr. Yashar pointed out that in his series there were no survivors if the mediastinal nodes were involved. Checking back, I find that we resected the lung in 48 patients who had recurrent nerve paralysis, and the five-year survival in that group was 8%. We resected the lung in the presence of pleural involvement in 74 patients with a five-year survival of 7%. We resected the lung with chest wall in 86 patients; the five-year survival in that group was 8%. Dr. Rasmussen mentioned prevention. I estimate that 3,500 of the 3,808 patients I have seen would never have developed their cancer if someone had convinced them when they were young not to have started smoking. I want to congratulate the members of this Society on their smoking habits now. I have been counting, and I have seen only 12 smokers; all the others are setting a good example. VOL. 20, NO. 5, NOVEMBER, 1975
519
P
rimary cancer of the lung was rare and always fatal before the early thirties. It is now common a n d selectively curable. A dismal picture, however, has been presented by most publications. They headline low overall cure rates, and even reports on results of mass screening programs are gloomy. During the past decade the majority of papers have focused on palliative programs of irradiation or chemotherapy. They advocate exhaustive attempts to obtain tissue or cells from sputum o r by extraction through endoscopic or direct needle aspiration. Many current articles stress the importance of avoiding operation. Our experience suggests that such intervention is necessary not only to establish the existence of the cancer in its early stages, but also to determine its extent in the advanced case.
Clinical Material and Methods Over more than four decades, 3,938 patients suspected of having primary malignancy of the lung have been observed and 3,808 have had a tissue diagnosis. The policy of management has largely been influenced by the senior author, and shifts in staff have been minimal: a total of 8 surgeons have been involved. Only histologically verified cases are the subject of this report. All bronchial adenomas, 175 in number, are excluded. T h e reference background is unique. In the first From the Overholt Thoracic Clinic, New England Baptist and New England Deaconess Hospitals, Boston, Mass. Presented at the Eleventh Annual Meeting of The Society of Thoracic Surgeons, Montreal, Que., Canada, Jan. 20-22, 1975. Address reprint requests to Dr. Overholt, Overholt Thoracic Clinic, 135 Francis St., Boston, Mass. 02215.
VOL. 20, NO. 5, NOVEMBER, 1975
511
OVERHOLT, NEPTUNE, AND ASHRAF three decades, many patients were referred from older populations of tuberculosis sanatoria. Twenty-four percent of those who underwent resection had diabetes or an associated serious disease. All patients with a verified diagnosis (many of whom were terminal) seen in consultation with other surgeons are included. Sixty-nine patients had a second resection, and 1 of these had a third. There were 37 ipsilateral and 32 bilateral operations. In 4 instances the second resection was either a pneumonectomy in a patient who had previously undergone resection of a portion of the contralateral lung or segmental resection of the remaining lung in a patient previously treated by pneumonectomy. Therefore this series of patients, like any other, has built-in variables as to selectivity. The first patient, treated in April of 1932 [2], was a 59-year-old man with severe chest pain. He had a necrotic epidermoid lesion the size of a grapefruit in the right lower lobe, and a palliative lobectomy was done. The lesion had invaded the pericardium. Convalescencewas uneventful; the hospital stay was 17 days. He received postoperative irradiation and spent a comfortable 10 months, then died suddenly while eating breakfast. Since then, of all the patients seen, 1,182 were rejected for operation, 57% were explored, and 48% of the total had resection. This summary focuses largely on the last group, numbering 1,848 patients. In the entire series the ages covered a spread from 15 to 93 years (average, 59) with the majority of patients between 45 and 70 years. The youngest resected patient was 15 and the oldest, 86 years. Clinical staging in retrospect according to the American Joint Committee recommendations of September, 1973 [ 11is impossible. Patients have simply been divided into two groups: those whose cancer was strictly localized to the primary site and those with evidence of spread either to a lymph node or by direct extrapulmonary extension. Of the resections performed, 1,162 were considered palliative and 686 curative. The resected tumors were adenocarcinoma in 17%of the patients, epidermoid in 4476,undifferentiated in 28%, .small cell in 5%, and other cell types in 6%. With time, the extent of resection has become more limited. Two factors are responsible: lesion size and acceptance of more patients with a low pulmonary reserve (Fig. 1).The Table shows operative mortality in the early and late periods of our experience according to the extent of resection as well as the objective of the procedure, i.e., for cure or palliation or only for exploration. During the first FIG. 1 . The resection experience is divided f o r convenience into four periods of ten years each with two extra years added to thefirst period. The change in the extent of resection performed is illustrated by decades. ( s e g = segmentectomy; lob = lobectomy; pn = pneumonectomy.)
%
25 I
512
T H E ANNALS OF THORACIC SURGERY
Primary Cancer Ofthe Lung OPERATIVE MORTALITY IN 1,848 RESECTIONS FOR LUNG CANCER
Mortality (%)
Procedure
Early
(1932- 1954)
Late (1955- 1974)
EXTENT OF RESECTION
Segmentectomy Lobectomy
Pneumonectomy
25 11
17
3
5 14
GOAL OF OPERATION ~~
Exploration only Curative resection Palliative resection
4 11 14
3 5 12
decade of this experience, operations were performed without the benefit of a blood bank, antibiotics, steroids, ventilatory assist devices, electronics, or a recovery room or intensive care unit. Blood from a bank was available in 1941, then penicillin in 1944. The other refinements came during the last half of this experience. All but 29 of the patients who underwent resection have been followed to the time of death or to survival for more than five years. Data on five-year survival when follow-up was possible revealed that 95% of the 1,960 untreated patients died in the first year after diagnosis (average survival, 4.2 months for patients rejected for operation and 6.1 months for those who were only explored). Results in treated patients are recorded according to the histology (Fig. 2), extent of resection (Fig. 3A), palliative or curative resection (Fig. 3B), and combinations of therapy (Fig. 4). Either alone or in combination, 747 patients received irradiation and 234 had chemotherapy.
Comment Out of all this experience, many lessons have been learned, but comment will be limited to three important points: the circumstances of discovery, simplification of the approach to the localized lesion, and development of a more direct approach to disseminated disease. In our series, only 3 of the cured patients were apprehended at a time when the chest films were normal. These patients all had symptoms that led to sputum investigation or bronchoscopy with biopsy, lavage, or brush. All the asymptomatic patients were discovered by radiological survey. Massive population screening for early cancer by cytological examination of bronchial secretions does not appear to be practical. In the past, the medical profession -including many radiologists - has not taken small, even minute areas of abnormal uncalcified density seriously enough VOL. 20, NO. 5, NOVEMBER, 1975
513
OVERHOLT, NEPTUNE, AND ASHRAF % 100
75 5 0
2 5.
5Yr
FIG. 2. Five-year survival according to histological type in patients who underwent resection shows that the outcome with adenocarcinoma (adeno) is 33 slightly more favorable than that with efndermoid 26 carcinoma (epid). The one-, two-, and three-year 22 survival exceeded by three times that of the untreated group. There were no long-term survivors among the 0 small or oat cell group. (undiff = undifferentiated carcinoma.) Sr
in terms of their cancer potential. T h e interval between the initial discovery film or onset of symptoms and the surgical investigation has been too long. This interval in the first decade of our experience was 18 months; in the second, 10 months; in the third, 9 months; and in the last, 5 months. Three paradoxes exist which indicate the need for a more direct course of management: (1) the fewer the symptoms and abnormal signs, the greater the likelihood that the substance producing a roentgenographic shadow is a cancer; (2) curability is inversely proportional to the number of symptoms and signs; and (3) the easier it is to obtain specimens for verification, the worse the prognosis. Therefore, if a person harbors a cancer within the lung, the most fortunate circumstance for him would be to have the shadow of the cancer discovered by chest roentgenogram at a time when he is apparently in robust health and free from symptoms or abnormal physical signs, with laboratory studies all within normal limits, the lesion not visible bronchoscopically, and the cytology negative regardless of how the specimen was obtained. The highest cure rate is attained in this set of circumstances. For isolated lesions in good-risk patients the approach can be direct. If surgical excision is indicated, regardless of a positive or negative sputum, bronchoscopy, biopsy, or brush, such investigations become superfluous. Needle aspiration of a suspicious area in the lung is also inconclusive and in addition is
7 5.
66
50.
40
2512
FIG. ?. Survival in 1,582 patients based on (A) extent of resection and (B) palliative or curative resection. The less extensive the resection, the better the survival. Note that life was polonged beyond one year for 40% of the resected patients (as compared with 5 %f o r those untreated). One hundred twenty-nine cases were discovered by survey; the survival rate was highest in this group. (seg = .regmentectomy; lob = lobectomy; pn = pneumonectomy.)
514
T H E ANNALS OF THORACIC SURGERY
Primary Cancer Ofthe Lung
15
FIG. 4 . ( A ) Palliative treatment consisted of a combination of chemotherapy and irradiation in 7 1 patients. In 2 7 of these it was possible to do a resection also. The triple combination produced better results in terms of prolongation of lfe. ( B ) Supervoltage irradiation was given to 478 patients, 200 of whom also had a resection. (C) One hundred j$y-seven patients were treated @ chemotherapy, and of these 77 also had a resection. Obviously, those who could be resected had less advanced disease, and results were expected to be better. However, in those patients treated only by chemotherapy, the survival curve ZS almost identical to that for the untreated group.
hazardous. Preoperative investigation should focus on cardiopulmonary reserve more than on ways to obtain tissue for verification. In patients with small, localized lesions who are asymptomatic or relatively so, little preparation is indicated. In the first decade of our experience, the relationship of smoking to chemical bronchitis and emphysema was less well appreciated than it is today. The principal etiological factor that initiates the cancer also diffusely damages the entire lung. Fortunately, chemical bronchitis is reversible. Cessation of smoke inhalation is the only preparation necessary for most patients in this category. The issue in disseminated cancer should also be settled directly and positively. Isotopic screening of skeleton or liver may give presumptive evidence of dissemination. Bone marrow aspiration may be indicated. The routine in many clinics is a series of studies performed several days apart. The investigation proceeds from one to the next test until positive tissue or cells are obtained. Bronchoscopy usually comes first, then cervical node biopsy, then percutaneous needle aspiration of the lesion, and finally mediastinoscopy or direct mediastinal exploration. Multiple iqyestigations can often be scheduled for the same day with a pathologist present for frozen section. In recent years, the first approach has often been directly to the liver or retroperitoneal nodes by laparotomy. If involvement is found below the diaphragm, either operation or irradiation of the lung would be futile gestures as far as hope for cure is concerned. At this point in time, surgical removal continues to provide the best hope for prolongation of life and possible cure. It is therefore imperative to focus on factors
VOL. 20, NO. 5 , NOVEMBER, 1975
515
OVERHOLT, NEPTUNE, AND ASHRAF which are basic to the surgical approach. The surgeon is responsible for determining the identity as well as extent of the lesion. Both should be accomplished in the most direct and absolute way possible. As one perfects, one simplifies. For the potential victim with a localized lesion and an adequate pulmonary reserve, histological verification and total biopsy to determine the extent of the cancer settle both issues and affect a cure simultaneously. Multitudinous preoperative investigations then become superfluous. The surgeon must also settle the issue when the lesion has extended beyond the primary site. Again, the most direct approach for verification and determination of extent should be taken. Hopelessness should be established early so that the patient can be spared suffering, expense, and useless days in a hospital if the terminal period is approaching. This experience has taught us to be more conservative in the extent of resection for small localized lesions. The value of palliative resection is now better appreciated in terms of prolongation of life, its quality, and, for some, the actuality of cure.
References 1. American Joint Committee for Cancer Staging and End Results Reporting. A Clinical Staging System for Carcinoma of the Lung, Chicago, 1973. 2. Overholt, R. H. Primary carcinoma of the lung: Surgical extirpation. A m ] Surg 22:181, 1933.
Discussion DR.J. GORDON SCANNELL (Boston, Mass.): Even though both Dr. Shields and Dr. Overholt made their manuscripts available to me in advance, it is clearly a challenge to discuss these two papers on cancer of the lung when one realizes that some 6,000 patients with a “surgical” type of disease are involved. Since one study is collaborative and the second is a personal experience, they have to be dealt with separately. The conclusions reported in the first paper by Shields and his group are an interesting by-product of a study designed for another purpose, namely, the evaluation of adjuvant chemotherapy. There is therefore an important element of preselection in it: only those who could be resected are considered. Under the circumstances, it is really not too surprising that cell type in itself is not a very meaningful determinant, particularly if one excludes oat cell carcinoma. I assume that bronchial adenoma has been omitted - or at least not tucked away into any of the groups -for though it may indeed be a malignant disease, it is not cancer of the lung in the ordinary clinical sense. In the same way, I would like to be sure that bronchiolar and alveolar cell carcinoma are not included. These remarks are not made in the spirit of captious criticism. I think the authors are very much to be commended for reporting their results in terms of staging of the disease. I do find it very interesting that they can distinguish so clearly between hilar and mediastinal involvement. Is there a mythical boundary that perhaps some of us have difficulty seeing? Or is it, in effect, a quantitative description of the degree of lymphatic involvement? I confess to a certain disappointment that postoperative irradiation in the presence of positive nodes was not included in the study. This, to me, still remains an important unanswered question in the surgical management of cancer of the lung, although I recognize that very real steps are being made by the Michigan group in this field. Personally, I have the clinical conviction (or prejudice) that postoperative radiation therapy in the presence of positive nodes, whether squamous or adenocarcinoma, has a real place in the management of the patient, and probably also of the cancer. Now that mediastinoscopy permits more accurate preoperative staging, the relative merit of postoperative versus preoperative irradiation, which still is a somewhat unanswered question, becomes important. 516
THE ANNALS OF THORACIC SURGERY
Primary Cancer of the Lung You really have to change gears to go from a collaborative study to the experience of a surgeon who has made cancer of the lung one of the personal challenges of his long, distinguished, and active career. It is all too easy - as I think Dr. Overholt implies -to depersonalize cancer of the lung, and I sense some impatience with too elaborate efforts to reconnoiter the field before launching an attack. One cannot help agreeing with a great deal of this, but I do feel that if we are going to make progress in defining the place of surgery, we will have to do more than divide operations into the categories of those done for “cure” and those for “palliation.” Both terms are in the eye of the beholder, in this case the operating surgeon. What is “palliative” is a matter of judgment, as shown by the fact that some 13% were actual survivors. But to evaluate the results of operation only in terms of cure is probably too harsh a judgment - again, as Dr. Overholt has implied. I would like to think (and I admit to a certain amount of chauvinism in this) that the reason limited anatomical resection -that is, lobectomy - became acceptable is attributable in no small measure to the ideas of my former mentor and Dr. Overholt’s contemporary, Dr. Edward Churchill. Certainly, both have taught us to carry the battle to the enemy, and if I may pontificate for the moment, both have accepted St. Paul’s injunction: “If the trumpet give forth an uncertain sound, who will prepare for the battle?” DR. DONALD L. PAULSON (Dallas, Tex.): This is really a remarkable morning, and I am privileged to follow Dr. Overholt’s excellent classic paper covering more than forty years’ experience by discussing Dr. Shields’ paper. When we examine separately the influence of cell type on prognosis, we find a strong relationship, particularly in the case of well-differentiated epidermoid carcinoma in contrast to very undifferentiated small cell carcinoma. However, taking all the variables together, histological type is of secondary importance to the stage and extent of disease. The relationship of histological type and mediastinal lymph node metastasis is noteworthy, because in both anaplastic carcinoma and some adenocarcinomas there is a much higher incidence of detected involvement than in the epidermoid cell type. Once the nodes are involved, however, the prognosis is poor, though variable in its course. Much of the confusion regarding the meaning of mediastinal node involvement arises from a lack of precise delineation of the location and extent of nodes involved. The location is characterized as being ipsilateral or contralateral, high or low peritracheal, and tracheobronchial. Those patients with contralateral or high peritracheal spread are usually inoperable, and their life expectancy is less than two years. Only those with involvement of the tracheobronchial nodes may be considered suitable for operation. In our series of 251 resections in patients with mediastinal node involvement, only 15 survived five years. All of these were patients with ipsilateral tracheobronchial nodes involved intranodally. The extent of node involvement, i.e., intranodal or perinodal, is very relevant, as has been shown by Dr. Bjork. Bergh and Schersten found that the operative mortality for patients with perinodal metastases is about five times greater and the survival one-tenth as good as in those with intranodal metastases. Unfortunately, perinodal involvement predominates in all histological types, but ipsilateral intranodal growth is found in about 12% of all cases. The operative mortality corresponds well with the established five-year survival rate of less than 10% after resection in patients with mediastinal node metastases. We do not think patients with mediastinal node involvement are automatically inoperable, but that there are some patients better suited for operation, the favorable lesions being those in the presence of mediastinal metastases that are ipsilateral, low peritracheal, or tracheobronchial in location and intranodal in extent of involvement. DR.MARVINM. KIRSH(Ann Arbor, Mich.): Since 1955, superior mediastinal, subcarinal, peritracheal, and paraesophageal mediastinal node dissection has been carried out following pulmonary resection at the University of Michigan Medical Center. If these nodes were positive, the patients received postoperative radiation therapy. The overall five-year survival for the group undergoing curative resection has been around 30%. In contrast to Dr. Shields’study, our results suggest that histological cell type is an important factor in determining the prognosis in patients with bronchogenic carcinoma, especially if there is evidence of lymph node metastasis. As a group, patients with squamous cell carcinoma of the lung had a greater five-year survival than patients with VOL. 20, NO. 5, NOVEMBER, 1975
517
OVERHOLT, NEPTUNE, AND ASHRAF adenocarcinoma of the lung. Even in the absence of lymph node metastasis, patients with squamous cell carcinoma tended to survive longer than those with adenocarcinoma. The difference in survival between the two groups was more striking when there were hilar or mediastinal node metastases. The five-year survival rate in patients with squamous cell carcinoma and hilar metastases treated by pulmonary resection alone was 45%, and it was 29.5% in patients with mediastinal metastases treated by pulmonary resection, mediastinal dissection, and postoperative mediastinal irradiation. In contrast are the poor results in patients with adenocarcinoma and lymph node metastases. None of the patients who had adenocarcinoma with hilar metastases lived five years, and only 5.9% of those with adenocarcinoma of the lung and mediastinal node metastases survived five years with operation, node dissection, and irradiation. In addition, this study also suggests that in patients with squamous cell carcinoma of the lung, it is the presence and not necessarily the extent of metastases that determines the prognosis, provided that patients with squamous cell carcinoma of the lung and mediastinal metastases receive radiation therapy postoperatively. Because of the poor results we have achieved with adenocarcinoma of the lung and mediastinal metastases over the past ten years, we are currently reevaluating our approach for these individuals. DR. JAMES J. YASHAR(Providence, R.I.): During the six-year period from 1963 to 1969, 148 patients underwent thoracotomy for bronchogenic carcinoma. We had a resectability rate of 84%; this is in contrast to the 48% reported by Dr. Overholt. Several factors influenced the survival rate. First, the tumor size: with tumors less than 2 cm in diameter the five-year survival was 61%; for those between 3 and 4 cm it was 36%; and between 5 and 7 cm, survival was 27%. We had no survivors among the patients who had a tumor greater than 7 cm. Our overall five-year survival was 34%. Second, lymph node involvement exerts an unfavorable influence on lung cancer. With no node involvement the five-year survival was 44%; with interlobar involvement, 33%; and with hilar involvement, 25%. We had no five-year survivors among patients who had mediastinal node involvement. This is statistically significant. The third factor is histological type. Our data indicate that patients with epidermoid carcinoma stand a better chance of having a resectable lesion. The survival rate for squamous cell carcinoma was 35%; for undifferentiated, 30%; for adenocarcinoma, 40%; and for bronchiolar carcinoma, 50%. These differences are not statistically significant. We grouped patients with oat cell carcinoma in a different category since these tumors behave differently biologically; we had no survivors among the patients who had an oat cell carcinoma. I would like to ask Dr. Shields and Dr. Overholt whether, based on their experience, they think there is any relationship between tumor size and five-year survival. DR. R. A. RASMUSSEN (Grand Rapids, Mich.): Dr. Overholt has presented a 42-year experience with 3,900 cases of lung cancer, emphasizing a high incidence in the fifth and sixth decades with an increasing percentage of women and a continuing need for early surgical consideration in all cases. He obtained a 48% five-year survival in favorable cases. This is an encouraging report from an experienced surgeon. Too much bad news, gloom, and doom are expressed by inexperienced workers, many of whom seem obsessed with a positive tissue diagnosis even though their methods are costly, time consuming, inadequate, and often as dangerous as diagnostic thoracotomy, which also offers the opportunity to perform an appropriate definitive resection. We must reorient the profession. There is good news about lung cancer. We know the cause -that it is seldom seen in nonsmokers and is therefore largely preventable. Periodic chest roentgenograms in smokers offer early diagnosis and detection. Once the tumor is detected and proper surgical resection accomplished, there is hope of a cure with good survival. We must emphasize both the use of diagnostic thoracotomy instead of all the lesser diagnostic techniques and avoidance of dangerous decisions on operability by generalists, internists, and radiologists. We will then improve our accuracy. Presently, surgical therapy offers the only real hope for victims of lung cancer. We find nearly 50% five-year survival in the favorable resected patient, and often there is surprising survival even in unfavorable 518
THE ANNALS OF THORACIC SURGERY
Primary Cancer of the Lung cases. Careful follow-up is necessary to ensure that death is due to persisting lung cancer and not to some other cause such as heart disease or the postpneumonectomy state. The picture is reasonably bright. In 1963 we reported a Western Michigan experience with 813 patients, emphasizing that lesser resections were as curative as pneumonectomy. We now have approximately 2,500 cases and maintain this position. I would affirm Dr. Overholt’s report and hope that we heed his suggestions. DR. SHIELDS:First, bronchial adenomas were omitted from the studies throughout. Alveolar cell tumors were excluded in the initial studiesand then later on were placed in the category of adenocarcinoma, since that is exactly what they are. One of the things that impresses me in all discussion on lung carcinoma is that great differences in long-term survival are reported - from as high as 48% to as low as 8% despite the fact that we are talking about the same disease. This can stem from only three things: selectivity in patient population, selectivity and methodology of pathological examination, and very small numbers of patients being used to determine long-term survival. Many of the results with small numbers are not significant. One of the spin-offs of the various collaborative studies is that a large number of patients are selected without personal prejudice and an evaluation of treatment is more readily possible. Concerning a comment by Dr. Scannell, when one is doing a randomized study to evaluate the use of adjuvant cancer chemotherapy, one does not have the option to use postoperative irradiation even if the mediastinal nodes are positive. However, I agree that such a study should be done, and one of the surgical protocols of the Working Party for the Therapy of Lung Cancer has been set u p to evaluate irradiation when the mediastinal nodes are diseased. Only from such a prospective randomized study will the true value of postoperative radiation therapy be found. In the past it has been used on a selectivebasis, so that its true value is undetermined; but certainly this adjuvant must be looked into. I believe it is extremely difficult to talk about tumor size except with peripheral lesions, since in central lesions distal alterations in the parenchyma due to inflammatory changes behind the tumor itself make it difficult to evaluate the true size. In the peripheral pulmonary nodule study that was conducted initially by Dr. John Steele, size was noted to be important: patients whose lesions were less than 3.5 cm had a much better prognosis than those with larger lesions. Also in this study, a small number of patients with oat cell tumors underwent surgical resection. (We believe peripheral oat cell tumors that have no nodal involvement should be resected.) T h e five-year survival for peripheral oat cell lesions was essentially the same as for adenocarcinoma, squamous cell carcinoma, and the undifferentiated large cell types. My last comments concern terminology. I don’t think the word palliation should be used to describe a resection (even though mediastinal nodes are involved) if all known tumor is removed. I believe it should be classified instead as curative resection. When tumor is left behind, the resection should be called incomplete; use of the term palliative in this situation is inappropriate. DR. OVERHOLT:Dr. Scannell raised a question about the exclusion of adenomas. These were omitted; there were approximately 90 of them. Also, I resected the right lung of a patient in November, 1933, and our pathologist called the tumor an adenocarcinoma. Later the diagnosis was changed to bronchial adenoma, and this case is not included. Dr. Scannell also questioned bronchiolar and alveolar types. We grouped those under adenocarcinoma. Dr. Yashar pointed out that in his series there were no survivors if the mediastinal nodes were involved. Checking back, I find that we resected the lung in 48 patients who had recurrent nerve paralysis, and the five-year survival in that group was 8%. We resected the lung in the presence of pleural involvement in 74 patients with a five-year survival of 7%. We resected the lung with chest wall in 86 patients; the five-year survival in that group was 8%. Dr. Rasmussen mentioned prevention. I estimate that 3,500 of the 3,808 patients I have seen would never have developed their cancer if someone had convinced them when they were young not to have started smoking. I want to congratulate the members of this Society on their smoking habits now. I have been counting, and I have seen only 12 smokers; all the others are setting a good example. VOL. 20, NO. 5, NOVEMBER, 1975
519