Long-Term CT Follow-Up in 40 Non-HIV Immunocompromised Patients with Invasive Pulmonary Aspergillosis: Kinetics of CT Morphology and Correlation with Clinical Findings and Outcome

Harald Brodoefel; Monika Vogel; Holger Hebart; Hermann Einsele; Reinhard Vonthein; Claus Claussen; Marius Horger


Am J Roentgenol. 2006;187(2):404-413. 

In This Article


Three hundred ten serial CT investigations of 40 patients (range, 2-20; mean, 7.7 per patient) were reviewed retrospectively over a mean follow-up of 166 days (range, 5-841 days). At the time of the initial chest CT, which was concomitant to the primary diagnosis of IPA, 26 (65%) of the patients were neutropenic. Twenty-three (57.5%) had previously undergone HSCT, and 12 (30%) were being treated with steroids for severe GVHD. Clinical signs of IPA such as fever or lower respiratory symptoms were present in all patients. Five patients had mild episodes of hemoptysis; in three patients, hemoptysis required surgical intervention.

Initial patterns of high-resolution CT scans were very small nodular in 25% (10/40), small nodular in 27.5% (11/40), large nodular in 45% (18/40), and consolidations in 15% (6/40). In 17.5% (7/40) of patients, nodules and consolidations coexisted. Two patients (5%) presented with additional extensive GGOs; in one patient (2.5%), a mixture of GGOs and a very small miliary pattern was the dominant feature. Multiple lesions were encountered in 28 patients (median, 3.5; mean, 5.2; range, 2-20 lesions), whereas in 11 patients IPA was diagnosed as the solitary focus.

An accompanying halo sign was observed in 87.5% (35/40) at day 1, and its median duration was 5 days (mean, 8; range, 1-30 days). The exact prevalence of the halo sign after day 1 was 62.5%, 37%, and 17.5% at days 4, 8, and 16, respectively. Conversely, the crescent sign showed an increasing prevalence with 5%, 10%, 25%, and 45% at days 4, 8, 16, and 32 postdiagnosis, respectively. It appeared at a median 13 days postdiagnosis (mean, 13 days; range, 1-32 days). In the course of the disease, 22 patients (55%) developed cavitations, five of them multiple with a maximum of three. The median time period until the appearance of cavities was 21 days (mean, 30; range, 6-93 days).

Figures 1A and 1B summarizes the characteristics of CT kinetics. At the day of IPA diagnosis, the median size of the largest reference lesion was 3.1 cm2 (mean, 7.3; range, 0.2-55 cm2), and the median lesion number was 3 (mean, 4; range, 1-20). Ninety percent (36/40) of the reference lesions showed a further increase in size after primary diagnosis, reaching a maximum median area of 12.5 cm2 (mean, 15.9; range, 0.4-55 cm2) at day 9 (median and mean; range, 1-36). This initial increase in area paralleled an only moderate rise in lesion numbers, the latter reaching a peak of 5.2 lesions (mean and median, 3; range, 1-20), also at day 9. After the increase in lesion size and number over the first 9 days, the size of the lesions showed a median plateau phase of 3.5 days until day 12.5 since IPA diagnosis (mean, 7 days until day 16; range, 1-87). However, contrary to the sizes of lesions, median numbers did not plateau, and within the same time, the interval decreased by 17% to 2.5 (mean, 3.7; range, 1-12). After this stable phase, most patients (32/40) showed a decrease of lesion size. In 62.5% (25/40) of patients, a reduction down to 50% of maximum size was registered 31 days after the initial diagnosis of IPA (median and mean, 40; range, 6-90). At the same time, the median number of lesions dropped to 2 (mean, 2.8).

Figure 1.

A and B, Kinetics of lesion size (A) and numbers (B). Median values and interquartile ranges are provided for lesion size, lesion number, and time (days). Time points (X) indicated are first diagnosis, first sight of maximal area, last sight of maximal area, time at halved maximal area, and time at complete radiologic remission. Mean values for these time points were 0, 9, 16, and 85.5 days, respectively. Ninety percent of 40 patients showed increase of lesion size after day of diagnosis; 62.5% had reduction down to 50% of maximum size, and 42.5% had complete radiologic remission.

Before cavitation, the increase of mean initial to maximum lesion size was noted to be 152%, from 8.5 cm2 to 21.4 cm2 (median, 4-15.7 cm2). In the absence of cavitation, however, the increase in size of lesions was only 61%, from 5.6 cm2 to 9 cm2 (median, 3.1-3.6 cm2). The initial increase in lesion area proved to be irrespective of the intensity of antifungal therapy (ratio of relative growths, 1.03; CI, 0.64-1.64). Median duration of medical therapy in our cohort was 58 days (range, 2-330 days).

According to the course of disease, all but two patients in our cohort were assigned to four groups ( Table 2 ). The remaining two patients showed a primarily stable form of IPA until they died of primary disease. Figures 2A,2B,2C,2D,3A,3B,3C,3D,4A,4B,4C, and 4D provide examples of disease courses.

Figure 2.

Remission with fibrotic residual lesion in 63-year-old man with acute lymphatic leukemia and neutropenia after intensive chemotherapy. A, On day invasive pulmonary aspergillosis (IPA) was diagnosed, axial CT scan showed multiple masses and nodules with halo sign. B, Eight days later, CT scan revealed confluence and increasing size of lesions. Combination of halo sign and extensive ground-glass opacity points to massive hemorrhage. C, Twenty days later, signs of active hemorrhage had disappeared and lesions had regressed to residual segmental or nonsegmental parenchymal consolidation. Right middle lobe showed signs of shrinkage. D, Finally, 90 days after initial diagnosis, two lesions had completely resolved, whereas largest mass has regressed to thin circular wall of fibrotic tissue without any further dynamics on follow-up CT scans. Patient survived IPA.

Figure 3.

Complete remission in 34-year-old man with acute lymphatic leukemia and aplasia after high-dose chemotherapy. A, On day of invasive pulmonary aspergillosis diagnosis, axial CT scan showed multiple nodules, one of three on given slice surrounded by halo of discrete hemorrhage. B, Eight days following initial diagnosis, lesions had increased in size and number and were still accompanied by halo sign. C, Fifteen days following initial diagnosis, halo sign had disappeared and all lesions were in decline. D, Forty-five days following initial diagnosis, chest CT showed complete radiologic remission.

Figure 4.

Postsurgical relapse in 54-year-old man with acute myelogenous leukemia and neutropenia after hematopoietic stem cell transplantation (HSCT). A, On day of invasive pulmonary aspergillosis diagnosis, axial CT scan showed solitary mass with discrete halo sign. Cavitation occurred at day 20 postdiagnosis and proved resistant to standard antifungal therapy. B, Two months postdiagnosis, walls of cavitation showed marked increase of thickness, and patient developed hemoptysis. For this reason, patient underwent lung resection 3 months postdiagnosis. C, Postoperative CT revealed bandlike fibrotic tissue and volume shrinkage but no signs of residual fungal infection. D, Three weeks postsurgery, chest CT showed relapse of fungal disease at site of resection with discrete crescent sign. Patient showed complete resolution months later related to prolonged antifungal therapy.

Seventeen patients (42.5%) had a complete radiologic remission of disease within a median 80 days (mean, 85.5; range, 10-240). Therefore, time until complete improvement was 2.5 times (CI, 1.007-6.3) as long in patients who showed cavitation of at least one of their lesions (median, 95; mean, 109.5; range, 25-240 days) when compared with those that did not (median, 50; mean, 51; range, 10-110 days). As evident from Figure 5 in our cohort, the median time until complete regression of all observed cavitary lesions was 90 days (mean, 84; range, 15-160 days) and significantly longer than the time until regression of lesions without cavitation formation (median, 30; mean, 38.9; range, 9-96; risk ratio, 0.5; CI, 0.31-0.76). Size of cavitary lesions did not affect the time until complete radiologic remission (slope, 4.1 d/cm2; CI, 2.9-11.1).

Figure 5.

Kaplan-Meier graph highlighting prolonged remission time of all observed cavitations in comparison with noncavitary lesions (risk ratio, 0.5; confidence interval, 0.31-0.76). Size of cavitations did not have significant effect on time to complete regression.

Among the 17 patients with full remission, nine had a complete normalization of findings on CT scans, seven had uncharacteristic fibrotic remnants, and one had two persistent thin-walled inactive cavities. Of note, all but two of those patients with fibrotic residual lesions had had cavitations in the course of their disease, whereas all but two patients without fibrotic remnants had not.

In nine patients (22%), complete remission was achieved through surgery within 30 days (median and mean, 77.3; range, 24-210 days) after IPA diagnosis with seven of the surgical resections performed early within the first 48 days. Indications for surgery were hemoptysis in three patients, removal of a solitary infectious focus before ablative chemotherapy for HSCT in four, before chemotherapy because of relapsing primary disease in one, and failure of medical therapy after HSCT in one patient.

Five of the 17 patients (29%) with remission without surgical intervention and four of the nine (44%) with surgical resection had a relapse of fungal disease ( Table 3 ). The latter were diagnosed 270 days after the start of disease (median and mean, 224; range, 10-556 days) and 140 days after complete resolution of radiologic signs of IPA (median and mean, 161; range, 20-480 days). In three patients, relapse occurred during ongoing therapy and in six patients, a median 130 days (range, 20-330 days) after cessation of antifungal treatment.

In those three patients with fungal relapse, in whom new lesions developed either at the location of a previous lesion (four lesions in two patients) or at the margins of a previous resection (one patient) or in postcavitary fibrotic tissue (one patient), relapse was interpreted as the reactivation of persistent microfoci; time until relapse was short in these cases: 20, 20, and 140 days after radiologic remission (GM, 38 days). For patients in whom recurrent lesions were found at anatomic sites that were initially unaffected and therefore regarded de novo rather than reactivated, however, the time to relapse was 4.4 times (CI, 0.98-19.6) as long after complete remission (GM, 167; range, 60-480 days). Similarly, the time until initial remission was 1.8 times (CI, 0.2-17) as long in patients with recurrence of foci as in those that showed true reinfection (GM, 75 and 41 days, respectively).

In contrast to patients with relapsed IPA, we found a smaller subset of five patients in our series (12.5%) who showed a bimodal course of disease. In these five patients, an initial phase of partial improvement was followed by progression in 144 days (median and mean, 142; range, 81-200) after initial diagnosis. This progression occurred in all but one case during standard antifungal therapy. In all five patients, relapse of IPA was accompanied by a halo sign and an increase in the size of existing nodules. A major increase in wall thickness also was noticed in four of six cavitary lesions. In addition, according to the definitions given earlier, we found a relapse of apparently cured lesions in four of five patients within this group.

Finally, seven patients (17.5%) in our series with a fulminant disease course at no point showed a decrease in number or size of lesions, despite adequate therapy. All these patients died a median of 20 days after the diagnosis of IPA (mean, 40; range, 5-64 days), and none developed either the crescent sign or cavitation.

In our series, 24 patients (60%) survived and 16 (40%) died, 8 (20%) of them because of IPA and at a median of 40 days postdiagnosis (mean, 70; range, 5-276 days). In detail, all seven patients with a fulminant disease course died because of IPA, two of them secondary to a cerebral manifestation and five because of IPA-related respiratory failure. Similarly, one of the patients with relapsing aspergillosis after initial remission died of IPA. All other deaths in our cohort were either related to pulmonary cytomegalovirus infection (n = 1) or refractory hematologic diseases (n = 7) and occurred 134 days postdiagnosis (mean, 158; range, 64-365 days).

Of all the parameters considered in our analysis, only the formation of cavitations had a statistically significant effect on outcome and time until 50% reduction of maximal lesion size or complete remission. Indeed, time to half maximal lesion size was 2.8 times (CI, 1.3-5.7) as long in the presence of cavitation as in its absence. Likewise, as evident from Figure 6, time until complete radiologic remission was 2.5 times (CI, 1.007-6.3) as long in patients with cavitary lesions as in those without. At the same time the chance to survive IPA was OR = 8.4 (CI, 1.07-176) times greater in patients with cavitary lesions than in those without ( Table 4 ). Initial or maximal lesion size and number, intensity and start of antifungal therapy, initial distribution of foci (unilateral vs bilateral), and absence or presence of HSCT, GVHD, aplasia, and primary disease did not have a considerable influence on either outcome or radiologic duration of disease.

Figure 6.

Box plot illustrating effect of cavitary lesions on radiologic duration of disease. Time until complete radiologic remission was 2.5 times longer in patients with cavitary lesions (confidence interval, 1.007-6.3).


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: