Association of lung carcinoma and tuberculosis.
Two hundred and twenty one consecutive cases of bronchogenic carcinomas were studied histologically for evidence of associated lesions. Seventy eight lesions were seen in 55 patients (24.8%). The most frequent was tuberculosis, seen in 29 patients. The next in frequency were scars, in 22, emphysema in 12 and thickened pleura in 7. Interstitial fibrosis, chronic bronchitis, lobar pneumonia and bronchiectasis, were other lesions, seen in very few cases. The types of carcinoma, in which associated lesions seen were, undifferentiated carcinoma (76%), adeno-carcinoma (56%), mixed tumor (37%), large cell anaplastic (25%), small cell anaplastic (23.7%) and epidermoid carcinoma in (5.44%). The incidence of tuberculous lesions in autopsies unassociated with tumor is 7%, as compared to 24.8% incidence of association with carcinoma; which is significant. There were seven scar cancers; with origin in tuberculous scars in two. The study indicates necessity of prospective study in this field.
Malignancy of lung is a major cause of death in developed countries and considerable attention is focussed all over the world on the genetics, the trigger and the mileu which favour the abnormal metaplasia and neoplasia at different anatomic sites.
With respect to human lungs, a large amount of data is available regarding its relationship with smoking and environmental pollutants. However, pulmonary infections, which may provide a mileu for the development of neoplasia, have not received enough attention, as their incidence is low in developed countries. Nevertheless, experimentally induced virus disease has been shown, to progress to neoplasia. The relation of pulmonary tuberculosis to neoplasia is alluded to in earlier literature,,,. However, there is a paucity of data on this aspect, especially from India. Consequently we decided to analyse the data available in the King Edward VII Memorial Hospital from this point of view.
Clinical and histopathological data of 221 successive cases of bronchogenic carcinoma comprising 125 biopsies and 96 autopsies were analysed. The tissues had been routinely fixed in neutral formalin and stained with haematoxylene - eosin. Special stains for keratin and mucin were done whenever required for histological typing.
The review of histopathological slides of 221 cases revealed 78 associated benign pulmonary lesions in 55 patients (24.88%). The ages of these patients ranged from 11-80 years, with a mean of 53 years. There were only seven women, with a mean age of 52.0 years. Of these lesions 93.6% were detected in the 96 necropsies, as larger number of sections were available for study.
The most frequent association with benign lung lesions was seen in undifferentiated carcinoma, (16/21 = 76%), followed by adenocarcinoma (45/81 = 56%). [Table:1]. The diagnosis of undifferentiated carcinoma was made whenever the cells did not show a definite pattern and were negative for mucus and keratin. These cases are likely to be either adenocarcinomas or epidermoid carcinomas, but most probably the former, as epidermoid carcinoma usually shows some stratification, which enables one to make a diagnosis. The lowest association with epidermoid carcinoma (5.44%) also substantiates the probability that most of the undifferentiated carcinomas were adenocarcinomas. If we group these two categories together, the incidence of associated lesions works out to be around 60%, which is very striking, in contrast to 5.44% with epidermoid carcinoma. Even if the two are not grouped together, the association in 56.25% of adenocarcinoma, which was the largest group, is significant. Small cell anaplastic carcinoma showed an incidence of 23.0%. The number of cases in the categories of large cell anaplastic carcinoma and mixed type are too small for comparison.
The details of the 78 associated benign lesions are given in [Table:2]. The most commonly associated lesion, seen in 29 out of 78 cases, (37.17%) was tuberculosis, either active or healing. In 16 cases, the tuberculous lesion showed significant scarring. In 5 cases, completely healed lesions or scars were seen some distance away from the tuberculous lesion. These could be presumed to be tuberculous too.
The next most frequent type of lesion (22 cases) was single or multiple areas of complete collagenisation, which did not give any clue to the preceding lesion. These scars could be the result of tuberculous or other granulomatous lesions, unresolved bronchopneumonia or infarcts, though the last usually result in linear scars on interstitial fibrosis. In 3 cases interstitial fibrosis was present, which could be the result of a multitude of processes including those mentioned above. In seven of these 22 cases, the lesion was classified as scar cancer. Tuberculous lesion with scarring was intimately associated with carcinoma in 2 cases. In 7 cases, there was thickened pleura, indicating healed pleural effusion, which in India is most frequently due to tuberculosis. Thus in a sizable number of the 29 cases (37.17%), there is a possibility of associated tuberculous lesion.
The association of emphysema in 12 cases cannot be further elaborated upon as the details about type and extent of emphysema are not available. Bronchicetasis and chronic bronchitis were noted only in 1 and 2 cases, respectively. The incidence of chronic bronchitis is very low in this series and is not truly reflected here as detection of chronic bronchitis requires detailed quantitative prospective study. The- association of these conditions and bronchogenic carcinoma therefore will not be further elaborated.
Co-existence of bronchogenic carcinoma and pulmonary tuberculosis was described by Bayle in 1810 but without definite histological proof. Penard in 1846 demonstrated it with definite histological evidence. Since then many workers in the latter part of 19th century and first part of 20th century have observed the association, but the opinion always remained divided on the point whether this association is fortuitous or whether it has any etiological basis. However, this association assumed significance with the emergence of the entity "scar cancers". It was postulated by Limas et al that during repair (following infection, inflammation or trauma) the epithelium is somewhat "restless" and consequently more inclined to undergo metaplasia. Further-more, carcinogens would preferentially concentrate at the site of hyperactivity to induce neoplastic change.
In the present series the coexistence of tuberculosis with bronchogenic carcinomas was detected in 29 out of 96 necropsy cases (30.2%). The incidence of tuberculosis amongst the general autopsies performed in the department does not exceed 7-8%. There is thus a significantly high incidence of tuberculosis in association with bronchogenic carcinoma. In another 3% of cases, there is a strong possibility of a tuberculous process being responsible.
In 7 cases in the present series, the carcinoma has been designated as scar cancer, amongst which, the fibrosis had a tuberculous aetiology in two cases.
In the typical scar carcinoma, which is a peripheral tumour, there is localised pleural scarring and puckering. The tumour shows a heavily anthracotic central scar with tumour tissue at the periphery,,. In many of these, evidence of vascular occlusion is seen. It is felt by some workers, that the scarring is a secondary phenomenon and not the cause of carcinoma. This however, is not applicable to scarring associated with tuberculosis, where it is a primary event. As mentioned before, no reports are available which distinguish these two types of scar cancers except that those associated with tuberculous scarring can be present anywhere. Since the present study was a retrospective one a detailed study with serial sectioning of cases with tuberculous scarring, could not be done. It is pertinent to emphasis that, though the association of tuberculosis and carcinoma was seen in 30.24% of cases definite linkage could be seen in only in two cases.
As regards the type of carcinoma, all the scar cancers were adenocarcinomas. In cases where there was no close proximity of the scar to the neoplasm, the most common type was also adenocarcinoma. The results of the present study are in confirmation with the observation of Yoneyama et al that epidermoid carcinoma is seen in an otherwise healthy lung, while adenocarcinoma, is more commonly associated with an already diseased lung.
Another interesting aspect of this problem is the linking of isoniazid with bronchogenic carcinoma,. In the present series there was some evidence of anti-tuberculous treatment in four cases, all of which showed active tuberculous lesions and carcinoma. However, none of these patients were tested slow or rapid inactivators of isoniazide. It is not possible, therefore, to draw any inference from these cases.
In conclusion, the present study shows a high incidence of tuberculosis in cases of bronchogenic carcinoma, as compared to cases without bronchogenic carcinoma. Though all types of bronchogenic carcinomas can be associated, there is more positive association with adenocarcinoma and negative association with epidermoid carcinoma. No definite relationship between the two can be established from this study, but it indicates the necessity for detailed studies of bronchial and alveolar epithelium in old tuberculous lesions, as well as, for collection of more epidemiologic data.
We wish to thank Dr PM Pai, Dean, King Edward Memorial Hospital, Mumbai, for permission to publish this data.