Update: The Radiographic Features of Pulmonary Tuberculosis pot

497 Update: The Radiographic Features of Pulmonary Tuberculosis John H. Woodring1 H. Mac Vandiviere2 Andrew M. Fried1 Marcus L. Dillon3 Terry D. Williams1 Irene G. Melvin2 Received April 1 5, 1985; accepted after revision November 4, 1985. 1 Department of Diagnostic Radiology, Albert B. Chandler Medical Center, University of Kentucky College of Medicine, 800 Rose St., Lexington, KY 40536-0084. Address reprint requests to J. H. Woodnng. 2Depment of Pediatrics, Albert B. Chandler Medical Center, University of Kentucky College of Medicine, Lexington, KY 40536-0084. 3Department of Cardiovascular and Thoracic Surgery, Albert B. Chandler Medical Center, Uni- versity of Kentucky College of Medicine, Lexington, KY 40536-0084. AJR 146:497-506, March 1986 0361 -803X/86/1 463-0497 C American Roentgen Ray Society Pulmonary tuberculosis produces a broad spectrum of radiographic abnormalities. During the primary phase of the disease these include pulmonary consolidation (50%), which often involves the middle or lower lobes or the anterior segment of an upper lobe; cavitation (29%) or pneumatocele formation(12%); segmental orlobar atelectasis(18%); pleural effusion (24%); hilar and mediastinal lymphadenopathy (35%); disseminated miliary disease (6%); and a normal chest radiograph (15%). During the postprimary phase of the disease, common abnormalities include exudative and/or fibroproductive parenchymal densities (100%), predominantly in the apical and posterior segments of the upper lobes (91%); cavitation (45%) with bronchogenic spread of disease (21%); marked fibrotic response in the lungs (29%); and pleural effusion, empyema, and fibrosis (18%, 4%, and 41%, respectively). Upper-lobe masslike lesions are seen occasionally (7%); spontaneous pneumothorax and intrathoracic lymphadenopathy are rare (5% each). Common causes of a missed diagnosis of tuberculosis are (1) failure to recognize hilar and mediastinal lymphadenopathy as a manifestation of primary disease in adults, (2) exclusion of tuberculosis because disease predominates in or is limited to the anterior segment of an upper lobe or the basilar segment of a lower lobe, (3) overlooking of minimal fibroproductive lesions or reporting them as inactive, (4) failure to recognize that an upper-lobe mass surrounded by satellite fibroproductive lesions might be tuberculous, and (5) failure to consider healed sequelae of primary disease or a positive purified protein derivative skin test as contributory to identifying the patient’s pulmonary disease. During the last decade, various authors have enumerated the “unusual” manifes- tations of pulmonary tuberculosis in the adult population [1-6]. These unusual manifestations usually have been implicated in the frequent failure of both radiolo- gist and clinician to recognize that tuberculosis could be the cause of an abnormal chest radiograph in patients who are subsequently, and rather surprisingly, proven to have tuberculosis. As Choyke et al. [7] indicate, these so-called unusual manifestations of tuberculosis in adults are actually typical of the disease: What is unusual is that they often represent primary disease occurring in the adult popula- tion. Although physicians are more familiar with the radiographic manifestations of postprimary tuberculosis than with those of primary disease, problems in radi- ographic diagnosis do occur. Minimal exudative or fibroproductive tuberculosis may be overlooked or, commonly, fibroproductive lesions may be assumed to be inactive even though the patient subsequently proves to be sputum-positive for Mycobacterium tuberculosis [1 , 5, 8]. Miller and MacGregor [1 ] reported several cases of advanced postprimary tuberculosis with extensive upper-lobe cavitation in which tuberculosis was not even considered as a diagnostic possibility! Because of the apparent widespread difficulty in recognizing the radiographic manifestations of pulmonary tuberculosis, we thought it would be beneficial to review the spectrum of radiographic abnormalities attributable to M. tuberculosis in a university medical center. We emphasize the radiographic features of primary 498 WOODRING ET AL. AGE IN YEARS Fig. 2 Primary lymphohematogenous dissemination with miliary pattern in 43-year-old man. AJR:146, March 1986 o PRIMARY TB  POST- PRIMARY TB Fig. 1 Prevalence of primary and postprimary tuberculosis by age in 90 patients. and postprimary disease, similarities and differences between the two, and the recognition of certain patterns that should lead to suspicion of pulmonary tuberculosis. The pathogen- esis of pulmonary tuberculosis is also discussed. Materials and Methods The chest radiographs and medical records of patients diagnosed as having clinically active M. tuberculosis were reviewed. Between 1974 and 1984, 107 patients with active pulmonary tuberculosis were documented at the University of Kentucky Medical Center. Of these, 17 patients had insufficient information in their medical records to determine whether they had primary or postprimary disease with M. tuberculosis. These 17 patients were excluded from the study. The Centers for Disease Control and American Thoracic Society criteria for confirmation of tuberculosis were met in the 90 remaining cases, that is, (1) either bacteriologic confirmation of current disease or both TABLE 1: Location of Pulmonary Consolidation in 73 Patients with Pulmonary Tuberculosis Site of consoldabon Type of Pul monary Tuberculosis Primary (n=17) Postpnmary (n=56) Right upper lobe: Anterior 4 18 Apical Posterior 0 1 41 31 Right middle lobe Right lower lobe: Superior Basal 2 1 3 10 13 16 Left upper lobe: Anterior 2 16 Apical Posterior 1 2 33 30 Lingula Left lower lobe: 2 14 Superior Basal Totals 2 4 14 10 24 246 a significant skin reaction to the purified protein derivative (PPD) and clinical and/or radiographic evidence of current disease, and (2) radiographic improvement after administration of two or more anti- tuberculous drugs [9, 10]. Primary tuberculosis was documented in 34 (38%) of the 90 patients on the basis ofthe following criteria: (1) recently documented conversion of the PPD skin test coupled with a positive culture of M. tuberculosis from tissue or body fluids (24 patients); (2) recently documented conversion of the PPD skin test coupled with clinical and radiographic evidence of current disease although subsequent cultures for M. tuberculosis were negative (seven patients, all chil- dren); and (3) positive culture for M. tuberculosis with no prior history of tuberculosis, a previously negative PPD skin test, and a previously normal chest radiograph (three patients). At the time of illness these patients were anergic and the PPD skin test was nonreactive. Of the 34 patients with primary tuberculosis, 19 were female and 15 were male; the patients were between 6 months and 87 years of age (average, 24 years) (fig. 1). Of these 34 patients, 19 (56%) were 18 years of age or older. These 19 adults with primary tuberculosis comprised 25% of the 75 adults with active pulmonary tuberculosis in our total study population of 90 individuals. As shown in figure 1, four of the patients with primary tuberculosis were over 71 years of age. There was no evidence that these four patients had ever had tuberculosis or a positive PPD skin test previously to suggest that the booster effect was the cause of the recent PPD conversion. One of these four elderly patients was quite debilitated with chronic lymphocytic leukemia. Two of the other adults with primary tubercu- losis had diabetes mellitus (one case each ofjuvenile-onset and adult- onset diabetes). Postprimary tuberculosis was documented in 56 (62%) of the 90 patients on the basis of the following criteria: (1) positive culture for M. tuberculosis obtained on one or more admissions to the hospital (all 56 patients), either from the sputum (53 cases), from material obtained from fiberoptic bronchoscopy (two), orfrom gastric aspirates (one); and (2) evidence of previous primary infection with tuberculosis (all 56 patients) based on a documented positive PPD skin test before the episode of current disease (39 patients), radiographic evidence of the previous primary infection (35), and/or a previously positive culture for M. tuberculosis without antituberculous treatment before C A . AJR:146, March 1986 PULMONARY TUBERCULOSIS 499 Fig. 3 Primary tuberculosis in 26-year-old man. A, Consolidation of the anterior segment of left-upper-lobe cavitation, left paratracheal lymphadenop- athy (arrow), and enlarged cardiac silhouette from pericardial effusion. Recently documented conversion of PPD skin test was believed to be noncontributory, and patient was treated with erythromycin for presumed Mycoplasma pneu- monia. B and C, 1 month later: spontaneous clearing of pericardial effusion. Cavitary disease in anterior segment of left upper lobe and left paratracheal lymphadenopathy (B, arrow) are more clearly seen. Sputum cultures obtained at this time grew Mycobacterium tuberculosis. the time of diagnosis (1 1). Of the 56 patients with postprimary tuberculosis, 36 were men and 20 were women; the patients were between 20 and 87 years of age (average, 51) (fig. 1). Five of these 56 patients had associated conditions that may have contributed to the development of postprimary tuberculosis, including chronic renal failure (two cases) and acute lymphocytic leukemia, chronic alcohol- ism, and pregnancy (one case each). The patients in our study were from central and eastern Kentucky. Although these patients are similar in many respects to those seen at other major medical centers, the majority came from small rural communities (less than 10,000 population) rather than urban inner- city areas. In fact, Kentucky is unique in that the case rate of tuberculosis is significantly greater in the nonurban population. The chest radiographs of the 90 patients were reviewed and evaluated for abnormalities of the lung parenchyma, airways, intra- thoracic lymph nodes, and pleura. Finally, the initial radiographic report and clinical diagnosis in each patient’s medical record was reviewed with attention directed to the degree of initial suspicion of tuberculous disease. The initial radiographic interpretations were prepared in conjunction by a radiology resident and board-certified diagnostic radiologist in all cases. Specific historical information con- cerning the status of the PPD skin test or a high degree of clinical suspicion of active tuberculosis was provided to the initial interpreters in less than one-half of cases. Potential causes of a nondiagnosis of tuberculosis were evaluated. Results Primary Tuberculosis A normal chest radiograph coupled with a positive culture for M. tuberculosis was found in five (15%) of the 34 patients ________ #{149}1 500 WOODRING ET AL. AJR:146, March 1986 Fig. 5 Primary tuberculosis in 3-year-old boy. Right hilar lymphadenopathy and thin-walled pneumatocele (arrows) in right lower lobe. Fig. 6 Primary tuberculosis in 4-year-old girl. Right lower-lobe collapse (black arrow) and shift of upper mediastinum to right (white arrow). with primary tuberculosis, who were clinically thought to have endobronchial disease. Two patients presented with a diffuse miliary pattern (fig. 2). The remaining 27 patients had pulmo- nary parenchymal consolidation, atelectasis, intrathoracic Fig. 4 Primary tuberculosis in 24-year-old female diabetic with recent conversion of PPD skin test. Multiple cavities with air-fluid levels (arrowheads) in superior segment of right lower lobe. lymphadenopathy, pleural effusion, and pericardial effusion as either isolated or combined findings. Pulmonary consolidation occurred in 17 patients (50%). In 13 patients a single area of pulmonary consolidation was encountered; in four patients multiple areas of parenchymal consolidation were present. Overall, 24 areas of consolidation were encountered in these 1 7 patients (table 1). The right and left lungs were equally affected; the mid and lower parts of the lungs were involved slightly more often than the upper lung zones; however, when the upper lobes were involved, anterior segment involvement was more common than in- volvement of the apical or posterior segments (table 1 , fig. 3). Cavitation had developed in five (29%) of the 1 7 patients with pulmonary consolidation (figs. 3 and 4); in one case air- fluid levels were noted. In two patients pneumatoceles were present (fig. 5). In seven of these patients pulmonary consol- idation was an isolated finding; however, in 10 patients con- solidation was associated with other findings including lymph- adenopathy, pleural effusion, pericardial effusion, and lobar atelectasis. Segmental or lobar atelectasis occurred in six patients (1 8%) (fig. 6). The anterior segment of the right upper lobe was involved in one case, the right middle lobe in two cases, and the right lower lobe in three cases. Pleural effusion was present in eight (24%) of the 34 patients (fig. 7). Effusion was unilateral in seven cases and bilateral in one. The size of the effusion was judged to be small in two cases, medium in one, and large in six according to the criteria of Roper and Waring [1 1]. Three patients presented with isolated pleural effusions; however, in five patients effusion was associated with parenchymal consoli- dation and/or lymphadenopathy. Pericardial effusion was seen in only one case (fig. 3). Echocardiography was not Fig. 7 Primary tuberculosis in 8-year-old girt. Large pleural effusion on Fig. 8 Primary tuberculosis in 13-year-old boy. Bilateral hilar, right para- right. tracheal, and aortic-pulmonic window lymphadenopathy (arrows) resembles sarcoidosis, lymphoma, or metastatic disease. AJR:146, March 1986 PULMONARY TUBERCULOSIS 501 performed routinely and therefore was not evaluated in this retrospective study. Hilar and/or mediastinal lymphadenopathy was present in 1 2 (35%) of the 34 patients (figs. 3, 5, and 8). Lymphadenop- athy was present in six of the 15 children and six of the 19 adults. Lymphadenopathy was limited to the hilum in six cases, involved both hilar and mediastinal nodes in three, and was limited to the mediastinum in three. In two patients lymphadenopathy was an isolated finding; however, in 10 instances lymphadenopathy was associated with pleural ef- fusion, consolidation, and/or atelectasis. The radiographic diagnosis was initially correct in only 11 (34%) of these patients with primary tuberculosis. Radi- ographic patterns most often diagnosed as tuberculosis in- cluded a miliary pattern, parenchymal consolidation accom- panied by lymphadenopathy in a child, and cavitation. Pat- terns least likely to elicit a diagnosis of tuberculosis included a normal chest radiograph, hilar and mediastinal lymphade- nopathy in an adult, and isolated findings of parenchymal consolidation, atelectasis, or pleural effusion. The clinical diagnosis was initially correct in 29 cases (85%) on the basis of the history and physical examination, chest radiograph report, and knowledge of a recent conversion of the PPD skin test. However, in five patients, all adults, tuber- culosis was initially excluded as a possible diagnosis because the chest radiograph was felt to be incompatible with the diagnosis of tuberculosis in an adult. Four of these patients demonstrated hilar and/or mediastinal lymphadenopathy and the fifth had consolidation of the anterior segment of the right upper lobe. Postprimary Tuberculosis Pulmonary parenchymal disease was present in all 56 patients with postprimary tuberculosis. In seven patients a single segment of lung was involved; in 49, two or more areas of the lungs were involved. Overall, 246 areas of pulmonary disease were present in these 56 patients (table 1). The apical and posterior segments of the upper lobes were the most commonly involved parts of the lungs (figs. 9 and 10). Involvement of unusual parts of the lung was common in postprimary tuberculosis (fig. 1 1). Twenty-four patients had disease involving the anterior segment of an upper lobe; 18 patients had involvement of the basilar segments of the lower lobes. However, in the vast majority of these cases, disease was present in more typical parts of the lung as well, for example, the apical and posterior parts of the upper lobes or superior segment of a lower lobe. In only one case each was postprimary tuberculosis isolated to the anterior segment of an upper lobe or basilar segment of a lower lobe. The pulmonary parenchymal disease was characterized as being exudative, fibroproductive (fibronodular and/or fibrocal- cific), and nodular. The most common finding (79%) was that of a mixture of exudative and fibroproductive lesions. In 18% the pulmonary disease was purely fibroproductive in nature, and in 3% the disease was purely exudative in nature. In 11 patients virtually all parts of both lungs were involved, and the disease was considered to be extensive. Lobar enlarge- ment was noted in one case. Large nodules or masslike densities were seen in four patients. These always occurred in the upper lobes and were always accompanied by other ‘ . I . . 4 ,, .    ,    114 502 WOODRING ET AL. AJR:146, March 1986 9 Fig. 9 Postprimary tuberculosis. Fibroproductive (fibronodular) densi- ties in apical segment of left upper lobe (arrows), initially interpreted as old, inactive scarring. Sputum culture obtained 1 year later grew Mycobacterium tuberculosis. Fig. 10 Postprimary tuberculosis in apical and posterior segments of both upper lobes in 36-year-old man. Blunting of right costophrenic angle (curved arrow) and calcified Ghon lesion in lingula (straight arrow) were residual findings from primary infection. /  p I r I Fig. 1 1 Postprimary tuberculosis in 77-year-old man. Homogeneous con- solidation of right middle and lower lobes and mottled consolidation of left lower lobe. Subtle fibroproductive lesions in apical segment of right upper lobe (straight arrow) and calcified Ghon lesion in left lower lobe (curved arrow). Bilateral apical pleural thickening. Fig. 12 Postprimary tuberculosis in 87-year-old woman. A 2-cm nonneo- plastic masslike density is seen associated with marked fibrosis of right upper lobe with upward retraction of right hilum and displacement of trachea to right. There is marked right apical pleural thickening. Fibroproductive lesions are present in basilar segments of right lower lobe and apical segment of left upper lobe. Calcified Ghon lesion in left lower lobe (arrowhead). Fig. 13 Postprimary tuberculosis in 20-year-old man. A 5-cm thick-walled cavity in posterior segment of right upper lobe contains air-fluid level (black arrow). Bronchogenic spread with acute tuberculous pneumonia involving basilar segments of right lower lobe, left upper lobe and lingula, and basilar segments of left lower lobe. White arrow indicates calcified aortic-pulmonic window node from previous primary infection. Fig. 14 Postprimary tuberculosis. Extensive fibroproductive and exudative densities with cavitation in apical and posterior segments of right upper lobe. Right hilar and paratracheal lymphadenopathy (straight arrows). Curved arrow indicates calcified Ghon lesion in right lower lobe. AJR:146, March 1986 PULMONARY TUBERCULOSIS 503 areas of fibroproductive or exudative densities (satellite le- sions) (fig. 12). Cavitation was common, being present in 25 (45%) of the 56 patients (fig. 13). In six cases a single cavity was present; in 19 cases there were multiple cavities. The cavity walls ranged from very thin and smooth to very thick and nodular. Air-fluid levels were present in five cases. Pneumatoceles were encountered in only one case. Bronchogenic spread of tuberculous material from upper-lobe cavities to other parts of the ipsilateral and/or contralateral lung was presumed to have occurred in 12 of the 56 patients (fig. 13) and often resulted in extensive abnormalities on the chest radiograph. A marked fibrotic response was evoked in the lungs of 16 patients (29%) (fig. 12). This was manifested by lobar atelec- tasis in nine patients, upward hilar retraction in 1 0, downward hilar retraction in one, tracheal deviation in seven, and ac- quired tracheomegaly in four. Pleural effusion occurred in 10 (1 8%) of the 56 patients with postprimary tuberculosis. Effusion was unilateral in eight cases and bilateral in two. The size of effusion was small in seven cases and medium in three. In two cases air-fluid levels were present in the pleural space, indicating bronchopleural fistula with empyema formation. Apical pleural thickening was present in 23 cases (41 %) (figs. 1 1, 12, and 14); it was unilateral in 14 cases and bilateral in nine. The extent of apical pleural thickening was usually 1 cm or less; in only three cases was the apical pleura over 1 cm in thickness. Sponta- neous pneumothorax was noted in three of 56 patients and was usually associated with extensive disease with multiple cavities. Hilar and/or mediastinal lymphadenopathy was present in three patients (5%) with postprimary tuberculosis (fig. 14). Lymph node groups involved included the right hilar, right paratracheal, and subcarinal nodes. The presence of intra- thoracic lymphadenopathy in these three cases accompanied rather extensive parenchymal disease with multiple areas of cavitation. By comparison with the lung disease, lymphade- nopathy was not a prominent feature of postprimary tuber- culosis in these three cases. Residual evidence of the original primary infection was apparent in 35 cases (63%). This included calcified or non- calcified Ghon lesions in 24 patients, calcified hilar and/or mediastinal lymph nodes in nine, and blunting of the lateral costophrenic angle in nine (figs. 10-14). The radiographic diagnosis was initially correct in 33 (59%) of these 56 patients. In 12 patients the radiograph was interpreted as abnormal but tuberculosis was not considered as a diagnosis, and in 11 patients the chest radiograph was interpreted as being normal (six cases) or as showing inactive lung scars (five cases). Common pitfalls in radiographic inter- pretation included the following: Fibroproductive lesions were either overlooked entirely or interpreted as being inactive; involvement of the anterior segments of the upper lobes or basal parts of the lower lobes was used as evidence against the diagnosis of tuberculosis; and large nodular or masslike densities were interpreted as being neoplastic. The clinical 504 WOODRING ET AL. AJR:146, March 1986 diagnosis was initially correct in only 31 cases (55%). In 12 patients the pulmonary symptoms were mild and nonspecific, and tuberculosis was not thought to be the cause; in addition, in several cases the chest radiographic abnormalities and positive PPD skin test were considered to be noncontributory to the present illness. In 13 patients the pulmonary symptoms were so slight that pulmonary disease was not clinically suspected. In all, there were 17 patients (30%) in which neither the radiologist nor clinician initially suspected tuber- culosis, resulting in a delay of 1 -3 years in the diagnosis. Discussion Primary tuberculosis is an airborne infection. The site of primary tuberculosis in the lungs reflects areas of greatest ventilation; the most common sites are the middle or lower lung zones or the anterior segment of an upper lobe [1 2]. The regional multiplication of organisms is followed by hilar and/ or mediastinal lymphadenopathy and the lymphohematoge- nous dissemination of organisms diffusely to the lungs and to numerous extrapulmonary sites [1 2]. During this time, most patients are either asymptomatic or develop mild to moderate systemic symptoms [1 2]. Tuberculin conversion occurs in 4- 8 weeks, at which time the patient will demonstrate a positive PPD skin test [12]. This corresponds to an activation of macrophage responsiveness, when activated macrophages become capable of containing the primary infection [1 2]. In 2%-6% of cases, the lymphohematogenous dissemination of massive numbers of viable organisms results in clinical and radiographic evidence of miliary tuberculosis [7, 1 2, 13]. This event is life-threatening and is characterized clinically by fever, weight loss, prostration, and, if untreated, death from respi- ratory failure and disseminated intravascular coagulation [12]. In about 10% of primary infections, acquired immunity is inadequate to contain the primary infection, and regional multiplication is followed by chronic, progressive parenchymal disease [1 2]. This phenomenon is distinguishable from post- primary disease only by documentation of a recent conversion of the PPD skin test [1 4]. In the remainder of cases the primary infection is contained but remains a significant risk for postprimary tuberculosis unless preventive treatment is given [1 5, 16]; memory lymphocytes become immunologically committed, and patients with primary infection generally retain both tuberculin reactivity and immunity from primary reinfec- tion for life [12]. Primary tuberculosis was once thought to be strictly a disease of childhood. During this century there has been a marked decrease in the incidence of tuberculosis in the United States as a result of careful public health measures and the advent of antituberculous chemotherapy. Because of this, only a small percentage of American adults have a positive PPD skin test, indicating that most adults today are suscep- tible to the development of primary tuberculosis [1 , 7]. We can be reasonably certain that 25% of the adults with pul- monary tuberculosis in our series had the primary form of the disease. Figure 1 shows three age-related peaks of highest prevalence. The highest prevalence of primary tuberculosis was noted in the 0-5-year age group; a second peak was noted in the 21 -30-year age group; and a third peak was noted in patients 71 years of age or older. The high prevalence of primary tuberculosis in young adults may be related to mobile, healthy, susceptible individuals living or working in crowded environments or under poor socioeconomic condi- tions [5]; primary tuberculosis in the elderly may be related to poor or crowded living conditions in addition to chronic or debilitating disease and immunosuppression [5]. Because of the recent increased prevalence of primary tuberculosis in adults, the term “childhood” tuberculosis to describe primary tuberculosis should be abandoned. Postprimary tuberculosis is the name given to a clinical and radiographic pattern of disease that correlates pathogeneti- cally with acquired hypersensitivity and immunity [1 3]. The careful observations and retrospective analyses of Stead [14, 17] and others [1 2] have led to the understanding that in 90% of cases, postprimary tuberculosis results from reactivation of a previously dormant primary iJ#{231}n; a minority of cases represent a continuation of the pnThary disease [13, 14]. Rarely, postprimary tuberculosis ‘is exogenous superinfection on an inactive or even active original infection-true reinfec- tion [13]. The early parenchymal infiltrate of postprimary tuberculosis results from the intensive inflammatory reaction in the hyper- sensitive host [1 8]. This may lead to caseous necrosis, which over time is apt to slough into a bronchus, leaving a cavity [1 8]. The most commonly involved parts of the lungs in postprimary disease are the apical and posterior segments of an upper lobe [1 , 5, 8, 13, 18-20]. Two main factors have been postulated to explain this common location of postpri- mary tuberculosis. The higher regional oxygen tensions of the apical and subapical parts of the upright lung (1 14 to 132 mm Hg) have been highly touted as being responsible for apical localization of postprimary disease. It is suggested that the demonstrated higher oxygen tensions in the apex of the upright lung produce increased mycobacterial virulence; the lower parts of lung where alveolar oxygen tension is 108 mm Hg or less are theorized to be quite resistant to the develop- ment of postprimary tuberculosis [1 8]. This concept may best be put to the test by observing populations dwelling at high altitude. Above 10,000 feet (3048 m) of elevation, alveolar oxygen tension is substantially less than 1 00 mm Hg; above 15,000 feet (4570 m), alveolar oxygen tension is well below 80 mm Hg, even in the extreme apex of the lung [181. Nevertheless, native permanent residents of the Peruvian highlands (1 0,300-1 4,300 feet [31 38-4357 m]) seem to be especially disposed to tuberculosis as compared to lowland natives (sea level) [1 8, 21]. A more likely explanation of the apical localization of post- primary tuberculosis is related to impaired clearance mech- anisms from poor lymph flow [1 8]. Pulmonary lymph flow is related to two factors: microvascular pressure and respiratory motion. In the erect lung, owing to gravitational effect, pul- monary perfusion and microvascular pressure are reduced in the apex relative to the middle and lower lung zones; as a result, apical lymph flow is also relatively deficient compared to the more dependent parts of the lung [1 8]. In addition, lymph flow in small lymphatic channels is directly related to AJR:146, March 1986 PULMONARY TUBERCULOSIS 505 respiratory excursion. Since the anterior rib cage moves with respiration but the posterior rib cage is relatively fixed, lymph flow is also restricted in the posterior parts of the lung [18]. In primary tuberculosis, lymphohematogenous dissemination results in the seeding of mycobacterial organisms to many or all parts of the lungs; the decreased lymph flow in the apical and posterior parts of the lungs would result in decreased clearance of organisms from these parts of the lung and would favor the development of postprimary tuberculosis in these locations [18]. Our study reveals a broad spectrum of radiographic abnor- malities in primary and postprimary tuberculosis not unlike those demonstrated in other reports on this subject. It also provides helpful insights into common causes for a misdi- agnosis of tuberculosis. In children, a clinical diagnosis of primary tuberculosis usu- ally is made when an abnormal chest radiograph is coupled with clinical evidence of pulmonary disease and a documented conversion of the PPD skin test. Certain radiographic pat- terns, such as parenchymal consolidation associated with hilar and/or mediastinal lymphadenopathy, or isolated hilar and/or mediastinal lymphadenopathy, in children elicit a fre- quent diagnosis of primary tuberculosis on the basis of the chest radiograph alone. In adults, these same findings often are viewed as unusual, atypical, or even incompatible with tuberculosis, and tuber- culin conversion may be viewed as a coincidental and unre- lated phenomenon. Four adults in our 34 cases of primary tuberculosis escaped initial radiographic and clinical diagnosis because hilar and/or mediastinal lymphadenopathy were con- sidered by the clinician and radiologist alike to be incompatible with “adult” tuberculosis. Hilar and mediastinal lymphadenop- athy are nonspecific findings that occur in bronchogenic car- cinoma, metastatic neoplasm, sarcoidosis, lymphoma, and fungal disease. But if the lymphadenopathy in these adults had been recognized as a possible manifestation of primary tuberculosis, the conversion of the PPD skin test might have been viewed as related to the patient’s illness, and a correct diagnosis of primary tuberculosis might have been made. In another adult patient, tuberculosis was excluded as a possi- bility because parenchymal consolidation was limited to the anterior segment of an upper lobe. While disease limited to the anterior segment of an upper lobe or basilar segment of a lower lobe is rare in postprimary tuberculosis and common in nontuberculous pneumonia, it should be remembered that these findings are rather common in primary tuberculosis. Postprimary disease in adults is often misdiagnosed also. The clinical symptoms often are minimal and the patient may report only a nonspecific “smoker’s cough” or no symptoms at all. In our series, 17 (30%) of 56 cases of active postprimary tuberculosis were missed by both the clinician and the radiol- ogist, resulting in a 1 -3-year delay in diagnosis on one or more hospital admissions. Incorrect appraisal of disease activity was the most com- mon cause of misdiagnosis. In local fibroproductive (fibro- nodular, fibrocalcific) tuberculosis the exudative lesion is grad- ually replaced by one or more sharply circumscribed shadows, which are often irregular and angular in contour [13]. Strand- ing toward the hilum occurs, and calcification in one or more of the nodules often appears [1 3]. This pattern of pulmonary abnormality is characteristic of granulomatous disease and is seldom, if ever, found in other bacterial infections. However, we found that it was this characteristic pattern fo parenchymal tuberculosis that caused the greatest difficulty in diagnosis. It is this stage of postprimary tuberculosis that is often over- looked entirely or is reported as “old scarring” or “no active disease.” This is incorrect, since many of these patients may be sputum-positive. Without films dating back at least 6 months, an accurate statement about the stability of a fibro- productive lesion cannot be made; all of these lesions should be reported as “possible tuberculosis or fungal disease, activ- ity indeterminant” [1 0, 22]. Even in patients with unchanging tuberculous scars, active tuberculous infection can still be present on sputum culture [1 ]. Therefore, these patients should be reported as “radiographically stable” rather than “inactive” [1 0, 22]. The term fibrotic, often used to describe this pattern of disease, is not recommended [1 0] because it implies inactivity. The second most common cause for a misdiagnosis of postprimary tuberculosis was involvement of an anterior seg- ment of an upper lobe or basilar segment of a lower lobe by parenchymal disease. It is a common misconception that involvement of these areas excludes tuberculosis. This is certainly not true of primary disease; moreover, in postprimary disease we found the anterior segments of the upper lobes to be involved in 46% of cases and the basilar segments of the lower lobes to be involved in 32% of cases. In most of these cases cavitary or fibroproductive disease was present in typical locations; many of these patients also showed residual radiographic evidence of the previous primary dis- ease-subtle but helpful adjunctive findings in arriving at a correct diagnosis. The third most common cause for misdiagnosis of postpri- mary disease was the presence of large nodular or masslike densities that were considered to be neoplastic. Typically, these were upper-lobe-predominant, rather than basilar, and were associated with fibroproductive lesions in typical loca- tions. As Palmer [1 9] states, “pulmonary tuberculosis continues to keep the physician humble.” Since the advent of antituber- culous chemotherapy 30 years ago, there has been a dramatic drop in the incidence of tuberculosis in this country. This has created a generation of American physicians who have little experience with it. Although the case rate of tuberculosis is relatively stable in many areas of the country, some areas are experiencing a marked increase in incidence. In addition, the recent increased prevalence of primary tuberculosis in adults has produced a confusing spectrum of disease seldom en- countered in adults in the past. Because of the infectious nature of tuberculosis, the morbidity and mortality of the natural course of the disease, and the excellent therapeutic results of antituberculous chemotherapy in most cases, there is good reason for all physicians to be sensitive to the contin- ued existence of this disease. The presence of healed Se- quelae of primary disease or a positive PPD should suggest that almost any radiographic abnormality could represent 506 WOODRING ET AL. AJR:146, March 1986 active tuberculosis. The recognition of primary disease in the adult and close scrutiny for and proper reporting of fibropro- ductive lesions could further improve the detection and diag- nosis of active tuberculosis. REFERENCES 1. Miller WT, MacGregor RR. Tuberculosis: frequency of unusual radiographic findings. AJR 1978;1 30:867-875 2. Lian CF, Robinson AE. Atypical radiographic appearance and clinical presentation of pulmonary tuberculosis. South Med J 1976;69: 131 7-1322 3. Amorosa JK, Smith PR, Cohen JR, Ramsey C, Lyons HA. Tuberculous mediastinal lymphadenitis in the adult. Radiology 1978;1 26:365-368 4. Liu C-I, Fields WR, Shaw C-I. Tuberculous mediastinal lymph- adenopathy in adults. Radiology 1978;1 26: 369-371 5. Miller WT. Tuberculosis in the adult. Postgrad Radio/ 1981; 1:147-167 6. Hadlock FP, Park 5K, Awe RJ, Rivera M. Unusual radiographic findings in adult pulmonary tuberculosis. AJR 1980;134: 1015- 1018 7. Choyke PL, Sostman HD, Curtis AM, et al. Adult-onset pulmo- nary tuberculosis. Radiology 1983;1 48:357-362 8. Woodring JH. The radiographic manifestations of pulmonary tuberculosis. J Ky Med Assoc 1984;82:17-23 9. Guide for completing reports of verified cases of tuberculosis. Atlanta, GA: Centers for Disease Control, Division of TB Control, 1984 10. Diagnostic standards and classification of tuberculosis and other mycobacterial diseases, 14th ed. Am Rev Respir Dis 1981; 123[suppl]: 343-358 11 . Roper WH, Waring JJ. Primary serofibrinous pleural effusion in military personnel. Am Rev Respir Dis 1955;71 :616-634 12. Geppert EF, Leff A. The pathogenesis of pulmonary and miliary tuberculosis. Arch Intern Med 1979;139: 1381 -1 383 13. Fraser RG, Pare JAP. Diagnosis of diseases of the chest, vol 2, 2d ed. Philadelphia: Saunders, 1978:731 -761 14. Stead WW, Kerby GR, Schlueter DP, Jordahl CW. The clinical spectrum of primary tuberculosis in adults: confusion with rein- fection in the pathogenesis of chronic tuberculosis. Ann Intern Med 1968;68: 731-745 15. Ferebee SH. Controlled chemoprophylaxis trials in tuberculosis, a general review. BibI Tuberc 1970;26:28-106 16. Dash LA, Comstock GW, Flynn JPG. Isoniazid preventive ther- apy: retrospect and prospect. Am Rev Respir Dis 1979; 1 19[suppl]:398 17. Stead WW. Pathogenesis of the sporadic case of tuberculosis. N EnglJ Med 1967;277:1008-1012 18. Goodwin RA, DesPrez RM. Apical localization of pulmonary tuberculosis, chronic pulmonary histoplasmosis, and progressive massive fibrosis of the lung. Chest 1983;83: 801-805 19. Palmer PES. Pulmonary tuberculosis-usual and unusual radi- ographic presentations. Semin Roentgenol 1979;1 4: 204-243 20. Jacobson HG, Shapiro JH. Pulmonary tuberculosis. Radio/ C/in North Am 1963;1 :411-427 21. Monge C Sr, Monge C Jr. High altitudediseases. Springfield, IL: Thomas, 1966 22. Diagnostic standards and classification of tuberculosis. New York: National Tuberculosis and Respiratory Disease Associa- tion, 1969 . infectious nature of tuberculosis, the morbidity and mortality of the natural course of the disease, and the excellent therapeutic results of antituberculous chemotherapy. inactive; involvement of the anterior segments of the upper lobes or basal parts of the lower lobes was used as evidence against the diagnosis of tuberculosis;