- Email: [email protected]
Imaging Features of Pulmonary Leukemic Infiltration With Comparison of Lymphoid and Myeloid Leukemias
ARTICLE IN PRESS Current Problems in Diagnostic Radiology 000 (2019) 1 4
Current Problems in Diagnostic Radiology journal homepage: www.cpdrjournal.com
Imaging Features of Pulmonary Leukemic Infiltration With Comparison of Lymphoid and Myeloid Leukemias Babina Gosangi, MBBS, MPH, Hiroto Hatabu, MD, PhD, Mark M. Hammer, MD* Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA
A B S T R A C T
Purpose: To describe the imaging findings of lung infiltration by leukemia and differential findings of lymphoid and myeloid leukemias. Materials and Methods: Through a search of electronic medical records from 2005 to 2017, we identified 21 patients with pathologically proven lung involvement by leukemia. Concurrent CT findings were analyzed by 2 chest radiologists in consensus for ground glass or consolidative opacities, septal thickening, bronchovascular bundle thickening, pulmonary nodules, pulmonary masses, and hilar and mediastinal lymphadenopathy. Results: There were 13 cases of lymphoid leukemias and 8 of myeloid leukemias. Nodules and masses were the most common imaging feature (n = 13, 62%). Bronchovascular bundle thickening and hilar lymphadenopathy were exclusively seen in lymphoid leukemias (P = 0.01 and P = 0.006). Centrilobular nodules were also exclusively seen in 3 patients with chronic lymphocytic leukemia. Conclusion: Lung infiltration by leukemia presents most commonly with nodules or masses, but interstitial abnormalities such as bronchovascular bundle thickening were seen as well. Radiologists should consider leukemic infiltration in the differential diagnosis for nodules, including centrilobular nodules, in these patients. © 2019 Elsevier Inc. All rights reserved.
Introduction
Material and Methods
Involvement of the lung by leukemia is frequently described at autopsy, ranging from 38% to upward of 60% of patients in some series.1-3 However, in many of these cases the leukemic infiltration does not manifest radiographically, and at least in some, it is likely to occur only at the end of life when imaging is not necessarily performed. Several prior series have described the imaging features of lung infiltration by leukemia.4-7 Common findings in these studies include interlobular septal thickening and bronchovascular bundle thickening as well as nodules. However, these prior studies were somewhat limited, in that one of these series focused only on lymphocytic leukemias, and the others had at most 11 patients. To our knowledge, no study has directly compared findings in myeloid and lymphoid leukemias. Knowledge of the protean manifestations of leukemic involvement of the lungs is important in order to consider this diagnosis in patients with leukemia and lung abnormalities. In this study, we describe the CT imaging features of pathologically proven lung involvement by leukemia. Additionally, we compare the different imaging findings in patients with lymphoid and myeloid leukemias.
Patient Selection
The authors did not receive any funding for this article. The authors have no financial disclosures. This study was approved by the Institutional Review Board. *Reprint requests: Mark M Hammer, MD, Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis Street, Boston, MA 02115. E-mail address: [email protected] (M.M. Hammer). https://doi.org/10.1067/j.cpradiol.2019.09.007 0363-0188/© 2019 Elsevier Inc. All rights reserved.
This study was approved by the Institutional Review board and conducted in accordance with HIPAA guidelines. We performed a retrospective review of patients with pathologically proven leukemia involvement of the lungs. To find patients, we searched the electronic medical records of 2 large, tertiary academic medical centers for pathology reports of lung biopsies (surgical or minimally invasive) or autopsies showing leukemic infiltration of the lungs, from January 2005 to December 2017. We only included cases that had a chest CT performed within a period of 1 month before or after the diagnosis was established by histopathology. Cases without concurrent CT scans and cases with pleural, mediastinal, or intrapulmonary lymph node involvement without lung parenchymal involvement were excluded from the study. Cases with concurrent leukemia and infection were excluded as well.
CT Image Review Cases were reviewed by 2 board-certified subspecialty thoracic radiologists in consensus. Images were analyzed for ground glass or consolidative opacities, septal thickening, bronchovascular bundle thickening, pulmonary nodules, pulmonary masses, pleural effusions, hilar and mediastinal lymphadenopathy, and cardiomegaly. If micronodules were present, their distribution pattern was recorded (centrilobular, perilymphatic, or random).
ARTICLE IN PRESS 2
B. Gosangi et al. / Current Problems in Diagnostic Radiology 00 (2019) 1 4
Statistical Analysis
TABLE 2 CT imaging features by type of leukemia
Data was stored in Microsoft Excel (Microsoft Corp, Redmond, WA) and analyzed with JMP Pro (v14, SAS Institute, Cary, NC). Fisher’s exact test was used to test associations between categorical variables, with a set to 0.05. Results Patient Clinical Characteristics A total of 21 cases of leukemia with pulmonary infiltration were identified (Table 1). Our study group included 12 men and 9 women with median age of 62 years (range 26-86 years). 10 patients had chronic lymphocytic leukemia, 5 had acute myeloid leukemia, 3 had acute lymphocytic leukemia, and 3 had chronic myelomonocytic leukemia. Median WBC count at the time of diagnosis was 17,000 cells/ml. Nine patients (43%) presented with fever or cough, and 4 patients (19%) complained of dyspnea at the time of presentation. Diagnoses were established through percutaneous CT-guided biopsy in 7 patients, transbronchial biopsy or bronchoalveolar lavage in 6 patients, surgical biopsy in 4 patients, and autopsy in 4 patients. Of the nonautopsy cases, the transbronchial biopsies were used in the case of diffuse lung disease (3 cases) and targeted to focal lung disease in 2 cases; in a single case, a lavage made the diagnosis of leukemia in the setting of multiple, diffusely distributed nodules and masses. All percutaneous biopsy cases were targeted to nodules or masses, as were 2 of the surgical biopsies, with the remaining 2 surgical biopsies used for diffuse lung disease.
Nodules Micronodules Masses Septal thickening Bronchovascular bundle thickening** Consolidation Diffuse ground glass Pleural effusions Hilar lymphadenopathy** Mediastinal lymphadenopathy* Cardiomegaly
Total (n = 21)
Lymphoid (n = 13)
Myeloid (n = 8)
10 (48%) 6 (29%) 5 (24%) 8 (38%) 7 (33%) 8 (38%) 5 (24%) 13 (62%) 8 (38%) 9 (43%) 4 (19%)
8 (62%) 5 (38%) 2 (15%) 4 (31%) 7 (54%) 6 (46%) 2 (15%) 8 (62%) 8 (62%) 8 (62%) 1 (8%)
2 (25%) 1 (13%) 3 (38%) 4 (50%) 0 (0%) 2 (25%) 3 (38%) 5 (63%) 0 (0%) 1 (13%) 3 (27%)
*P < 0.05. **P 0.01 for lymphoid vs myeloid.
CT Imaging Findings Imaging features with breakdown by type of leukemia are given in Table 2. Overall, nodules and masses were the most common lung finding (Fig. 1), present in 10 (48%) and 5 (24%) of patients, respectively (a total of 13 patients, or 62%, had either nodules or masses). In 6 patients, the nodules formed a micronodular pattern, with 3 cases of centrilobular nodules (all in patients with chronic lymphocytic leukemia, Fig. 2) and 2 of perilymphatic nodules. Septal and bronchovascular bundle thickening were common as well, present in 8 (38%) and 7 (33%) of patients, respectively (Fig. 3). Additional imaging features included consolidation in 8 patients (38%, Figs. 3 and 4) and diffuse ground glass opacities in 5 (24%).
FIG 1. A 87-year-old woman with AML. Axial CT image shows a left lower lobe mass (*). Percutaneous biopsy demonstrated leukemic infiltration. AML, acute myeloid leukemia.
TABLE 1 Patient and clinical characteristics N = 21 Age, yr (median, range) Male Type of leukemia AML ALL CLL CMML WBC at diagnosis, 1000*cells/ml (median, range) Method of diagnosis Autopsy Percutaneous biopsy Transbronchial biopsy Surgical biopsy Patient symptoms Cough and fever Dyspnea Fatigue
62 (26-86) 12 (57%) 5 (24%) 3 (14%) 10 (48%) 3 (14%) 17 (0.09-238) 4 (19%) 7 (33%) 6 (29%) 4 (19%) 9 (43%) 4 (19%) 1 (5%)
ALL, acute lymphocytic leukemia; AML, acute myeloid leukemia; CLL, chronic lymphocytic leukemia; CMML, chronic myelomonocytic leukemia.
FIG 2. A 71-year-old man with CLL. Axial CT image shows diffuse centrilobular (and tree-in-bud) nodules. Transbronchial biopsy demonstrated leukemic infiltration. CLL, chronic lymphocytic leukemia.
Comparing the myeloid and lymphoid leukemia cases, nodules were more commonly seen in lymphoid leukemias (62% vs 25%), but this did not reach statistical significance. Bronchovascular bundle thickening was exclusively seen in lymphoid leukemias (54% vs 0%,
ARTICLE IN PRESS B. Gosangi et al. / Current Problems in Diagnostic Radiology 00 (2019) 1 4
3
FIG 3. A 52-year-old woman with CLL. (a) Axial CT image shows right lower lobe consolidation with septal thickening (arrowheads). (b) Percutaneous biopsy in the prone position of the consolidation demonstrated leukemic infiltration. CLL, chronic lymphocytic leukemia.
FIG 4. A 50-year-old woman with ALL. (a) Axial CT image shows bilateral peribronchial consolidations (arrows). (b) Percutaneous biopsy of a lingular consolidation demonstrated leukemic infiltration. ALL, acute lymphocytic leukemia.
P = 0.01). The other pulmonary findings were similar across the leukemia types. In addition to the lung parenchymal findings, pleural effusions were present in 13 cases (62%), mediastinal lymphadenopathy in 9 cases (43%), hilar lymphadenopathy in 8 cases (38%), and cardiomegaly in 4 cases (19%). Hilar lymphadenopathy was only seen in lymphoid leukemias (62% vs 0%, P = 0.006). Discussion In summary, we reviewed 21 cases of leukemias involving the lung parenchyma, including acute, chronic, lymphoid, and myeloid subtypes, to evaluate their imaging appearance. We found that nodules and masses were the most common imaging feature (present in 62% of patients); less common imaging features included consolidation, septal and bronchovascular bundle thickening, and diffuse ground glass opacities. Bronchovascular bundle thickening and hilar lymphadenopathy were only seen in the lymphoid leukemias; nodules were also more common in these patients, although the difference was not statistically significant. Parenchymal nodules and masses are the most frequent finding in patients with leukemias in our study. This is consistent with work from Heyneman et al, who also found nodules to be the most common finding in their series.4 Of note, we saw 3 cases of centrilobular nodules, all in patients with lymphoid leukemias. While radiologists typically associate centrilobular nodules with infectious or inflammatory processes, these can rarely result from infiltration of leukemic cells along the bronchioles.8 Indeed, centrilobular nodules were found frequently in previous work focusing on lymphoid leukemias.7,9
We also found bronchovascular bundle thickening to be a unique feature in lymphoid leukemias and likely reflects the predilection of this malignancy for peribronchial lymphatics. This finding is consistent with prior work showing a high frequency of bronchovascular bundle thickening in lymphoid leukemias.8 Notably, the presence of septal thickening did not differ substantially between myeloid and lymphoid leukemias, although both presumably are related to perilymphatic infiltration. A substantial minority (43%) of patients in our series presented with clinical symptoms of infection (ie, fever and cough). With those symptoms in an immunocompromised patient, consolidations or centrilobular nodules are typically assumed to represent infections. However, recurrence or progression of the patient’s primary malignancy should always be kept in mind. In particular, in cases of patients with lymphoid leukemias presenting with centrilobular nodules, radiologists may suggest leukemic infiltration if the nodules increase concomitantly with worsening disease elsewhere and/or do not respond to antimicrobial therapy. In patients with any type of leukemia presenting with consolidations, radiologists should suggest this diagnosis if the abnormality persists or progresses despite appropriate antimicrobial therapy. The differential considerations for the latter scenario would also include aggressive atypical infections such as mucormycosis, and tissue sampling will likely be necessary to establish the diagnosis. It is important to distinguish infections from progression of leukemia, as these require substantially different management (ie, antifungals or antibiotics vs chemotherapy). Additionally, these 2 possibilities come with substantially different prognoses, as patients may recover from acute infections, while progression of
ARTICLE IN PRESS 4
B. Gosangi et al. / Current Problems in Diagnostic Radiology 00 (2019) 1 4
malignancy despite prior lines of chemotherapy portends refractory disease. Being a retrospective study, we are limited by inconsistent CT protocols and differing clinical management. Our sample size is small, although that is expected given the rarity of this entity. On the other hand, all cases in our series were proven pathologically to represent leukemic infiltration of the lungs. In conclusion, lung involvement by leukemia may manifest in a number of ways, ranging from focal lesions such as nodules, masses, and consolidations, to diffuse abnormalities such as septal and bronchovascular bundle thickening. Importantly, manifestations of consolidations and centrilobular nodules mimic infections, and in our series a substantial minority of patients presented with infectious symptoms. Radiologists should consider leukemic infiltration in cases where suspected infectious findings do not resolve with antibiotic or antifungal therapy.
References 1. Klatte EC, Yardley J, Smith EB, et al. The pulmonary manifestations and complications of leukemia. Am J Roentgenol Radium Ther Nucl Med 1963;89:598–609. 2. Maile CW, Moore AV, Ulreich S, et al. Chest radiographic—Pathologic correlation in adult leukemia patients. Invest Radiol 1983;18:495–9. 3. Barcos M, Lane W, Gomez GA, et al. An autopsy study of 1206 acute and chronic leukemias (1958 to 1982). Cancer 1987;60:827–37. 4. Heyneman LE, Johkoh T, Ward S, et al. Pulmonary leukemic infiltrates. Am J Roentgenol 2000;174:517–21. 5. Tanaka N, Matsumoto T, Miura G, et al. CT findings of leukemic pulmonary infiltration with pathologic correlation. Eur Radiol 2002;12:166–74. 6. Potenza L, Luppi M, Morselli M, et al. Leukaemic pulmonary infiltrates in adult acute myeloid leukaemia: A high-resolution computerized tomography study. Br J Haematol 2003;120:1058–61. 7. Okada F, Ando Y, Kondo Y, et al. Thoracic CT findings of adult T-cell leukemia or lymphoma. Am J Roentgenol 2004;182:761–7. 8. Trisolini R, Lazzari Agli L, Poletti V. Bronchiolocentric pulmonary involvement due to chronic lymphocytic leukemia. Haematologica 2000;85:1097. 9. Carmier D, Dartigeas C, Dartigeas C, et al. Serious bronchopulmonary involvement due to chronic lymphocytic leukaemia. Eur Respir Rev 2013;22:416–9.
Current Problems in Diagnostic Radiology journal homepage: www.cpdrjournal.com
Imaging Features of Pulmonary Leukemic Infiltration With Comparison of Lymphoid and Myeloid Leukemias Babina Gosangi, MBBS, MPH, Hiroto Hatabu, MD, PhD, Mark M. Hammer, MD* Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA
A B S T R A C T
Purpose: To describe the imaging findings of lung infiltration by leukemia and differential findings of lymphoid and myeloid leukemias. Materials and Methods: Through a search of electronic medical records from 2005 to 2017, we identified 21 patients with pathologically proven lung involvement by leukemia. Concurrent CT findings were analyzed by 2 chest radiologists in consensus for ground glass or consolidative opacities, septal thickening, bronchovascular bundle thickening, pulmonary nodules, pulmonary masses, and hilar and mediastinal lymphadenopathy. Results: There were 13 cases of lymphoid leukemias and 8 of myeloid leukemias. Nodules and masses were the most common imaging feature (n = 13, 62%). Bronchovascular bundle thickening and hilar lymphadenopathy were exclusively seen in lymphoid leukemias (P = 0.01 and P = 0.006). Centrilobular nodules were also exclusively seen in 3 patients with chronic lymphocytic leukemia. Conclusion: Lung infiltration by leukemia presents most commonly with nodules or masses, but interstitial abnormalities such as bronchovascular bundle thickening were seen as well. Radiologists should consider leukemic infiltration in the differential diagnosis for nodules, including centrilobular nodules, in these patients. © 2019 Elsevier Inc. All rights reserved.
Introduction
Material and Methods
Involvement of the lung by leukemia is frequently described at autopsy, ranging from 38% to upward of 60% of patients in some series.1-3 However, in many of these cases the leukemic infiltration does not manifest radiographically, and at least in some, it is likely to occur only at the end of life when imaging is not necessarily performed. Several prior series have described the imaging features of lung infiltration by leukemia.4-7 Common findings in these studies include interlobular septal thickening and bronchovascular bundle thickening as well as nodules. However, these prior studies were somewhat limited, in that one of these series focused only on lymphocytic leukemias, and the others had at most 11 patients. To our knowledge, no study has directly compared findings in myeloid and lymphoid leukemias. Knowledge of the protean manifestations of leukemic involvement of the lungs is important in order to consider this diagnosis in patients with leukemia and lung abnormalities. In this study, we describe the CT imaging features of pathologically proven lung involvement by leukemia. Additionally, we compare the different imaging findings in patients with lymphoid and myeloid leukemias.
Patient Selection
The authors did not receive any funding for this article. The authors have no financial disclosures. This study was approved by the Institutional Review Board. *Reprint requests: Mark M Hammer, MD, Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis Street, Boston, MA 02115. E-mail address: [email protected] (M.M. Hammer). https://doi.org/10.1067/j.cpradiol.2019.09.007 0363-0188/© 2019 Elsevier Inc. All rights reserved.
This study was approved by the Institutional Review board and conducted in accordance with HIPAA guidelines. We performed a retrospective review of patients with pathologically proven leukemia involvement of the lungs. To find patients, we searched the electronic medical records of 2 large, tertiary academic medical centers for pathology reports of lung biopsies (surgical or minimally invasive) or autopsies showing leukemic infiltration of the lungs, from January 2005 to December 2017. We only included cases that had a chest CT performed within a period of 1 month before or after the diagnosis was established by histopathology. Cases without concurrent CT scans and cases with pleural, mediastinal, or intrapulmonary lymph node involvement without lung parenchymal involvement were excluded from the study. Cases with concurrent leukemia and infection were excluded as well.
CT Image Review Cases were reviewed by 2 board-certified subspecialty thoracic radiologists in consensus. Images were analyzed for ground glass or consolidative opacities, septal thickening, bronchovascular bundle thickening, pulmonary nodules, pulmonary masses, pleural effusions, hilar and mediastinal lymphadenopathy, and cardiomegaly. If micronodules were present, their distribution pattern was recorded (centrilobular, perilymphatic, or random).
ARTICLE IN PRESS 2
B. Gosangi et al. / Current Problems in Diagnostic Radiology 00 (2019) 1 4
Statistical Analysis
TABLE 2 CT imaging features by type of leukemia
Data was stored in Microsoft Excel (Microsoft Corp, Redmond, WA) and analyzed with JMP Pro (v14, SAS Institute, Cary, NC). Fisher’s exact test was used to test associations between categorical variables, with a set to 0.05. Results Patient Clinical Characteristics A total of 21 cases of leukemia with pulmonary infiltration were identified (Table 1). Our study group included 12 men and 9 women with median age of 62 years (range 26-86 years). 10 patients had chronic lymphocytic leukemia, 5 had acute myeloid leukemia, 3 had acute lymphocytic leukemia, and 3 had chronic myelomonocytic leukemia. Median WBC count at the time of diagnosis was 17,000 cells/ml. Nine patients (43%) presented with fever or cough, and 4 patients (19%) complained of dyspnea at the time of presentation. Diagnoses were established through percutaneous CT-guided biopsy in 7 patients, transbronchial biopsy or bronchoalveolar lavage in 6 patients, surgical biopsy in 4 patients, and autopsy in 4 patients. Of the nonautopsy cases, the transbronchial biopsies were used in the case of diffuse lung disease (3 cases) and targeted to focal lung disease in 2 cases; in a single case, a lavage made the diagnosis of leukemia in the setting of multiple, diffusely distributed nodules and masses. All percutaneous biopsy cases were targeted to nodules or masses, as were 2 of the surgical biopsies, with the remaining 2 surgical biopsies used for diffuse lung disease.
Nodules Micronodules Masses Septal thickening Bronchovascular bundle thickening** Consolidation Diffuse ground glass Pleural effusions Hilar lymphadenopathy** Mediastinal lymphadenopathy* Cardiomegaly
Total (n = 21)
Lymphoid (n = 13)
Myeloid (n = 8)
10 (48%) 6 (29%) 5 (24%) 8 (38%) 7 (33%) 8 (38%) 5 (24%) 13 (62%) 8 (38%) 9 (43%) 4 (19%)
8 (62%) 5 (38%) 2 (15%) 4 (31%) 7 (54%) 6 (46%) 2 (15%) 8 (62%) 8 (62%) 8 (62%) 1 (8%)
2 (25%) 1 (13%) 3 (38%) 4 (50%) 0 (0%) 2 (25%) 3 (38%) 5 (63%) 0 (0%) 1 (13%) 3 (27%)
*P < 0.05. **P 0.01 for lymphoid vs myeloid.
CT Imaging Findings Imaging features with breakdown by type of leukemia are given in Table 2. Overall, nodules and masses were the most common lung finding (Fig. 1), present in 10 (48%) and 5 (24%) of patients, respectively (a total of 13 patients, or 62%, had either nodules or masses). In 6 patients, the nodules formed a micronodular pattern, with 3 cases of centrilobular nodules (all in patients with chronic lymphocytic leukemia, Fig. 2) and 2 of perilymphatic nodules. Septal and bronchovascular bundle thickening were common as well, present in 8 (38%) and 7 (33%) of patients, respectively (Fig. 3). Additional imaging features included consolidation in 8 patients (38%, Figs. 3 and 4) and diffuse ground glass opacities in 5 (24%).
FIG 1. A 87-year-old woman with AML. Axial CT image shows a left lower lobe mass (*). Percutaneous biopsy demonstrated leukemic infiltration. AML, acute myeloid leukemia.
TABLE 1 Patient and clinical characteristics N = 21 Age, yr (median, range) Male Type of leukemia AML ALL CLL CMML WBC at diagnosis, 1000*cells/ml (median, range) Method of diagnosis Autopsy Percutaneous biopsy Transbronchial biopsy Surgical biopsy Patient symptoms Cough and fever Dyspnea Fatigue
62 (26-86) 12 (57%) 5 (24%) 3 (14%) 10 (48%) 3 (14%) 17 (0.09-238) 4 (19%) 7 (33%) 6 (29%) 4 (19%) 9 (43%) 4 (19%) 1 (5%)
ALL, acute lymphocytic leukemia; AML, acute myeloid leukemia; CLL, chronic lymphocytic leukemia; CMML, chronic myelomonocytic leukemia.
FIG 2. A 71-year-old man with CLL. Axial CT image shows diffuse centrilobular (and tree-in-bud) nodules. Transbronchial biopsy demonstrated leukemic infiltration. CLL, chronic lymphocytic leukemia.
Comparing the myeloid and lymphoid leukemia cases, nodules were more commonly seen in lymphoid leukemias (62% vs 25%), but this did not reach statistical significance. Bronchovascular bundle thickening was exclusively seen in lymphoid leukemias (54% vs 0%,
ARTICLE IN PRESS B. Gosangi et al. / Current Problems in Diagnostic Radiology 00 (2019) 1 4
3
FIG 3. A 52-year-old woman with CLL. (a) Axial CT image shows right lower lobe consolidation with septal thickening (arrowheads). (b) Percutaneous biopsy in the prone position of the consolidation demonstrated leukemic infiltration. CLL, chronic lymphocytic leukemia.
FIG 4. A 50-year-old woman with ALL. (a) Axial CT image shows bilateral peribronchial consolidations (arrows). (b) Percutaneous biopsy of a lingular consolidation demonstrated leukemic infiltration. ALL, acute lymphocytic leukemia.
P = 0.01). The other pulmonary findings were similar across the leukemia types. In addition to the lung parenchymal findings, pleural effusions were present in 13 cases (62%), mediastinal lymphadenopathy in 9 cases (43%), hilar lymphadenopathy in 8 cases (38%), and cardiomegaly in 4 cases (19%). Hilar lymphadenopathy was only seen in lymphoid leukemias (62% vs 0%, P = 0.006). Discussion In summary, we reviewed 21 cases of leukemias involving the lung parenchyma, including acute, chronic, lymphoid, and myeloid subtypes, to evaluate their imaging appearance. We found that nodules and masses were the most common imaging feature (present in 62% of patients); less common imaging features included consolidation, septal and bronchovascular bundle thickening, and diffuse ground glass opacities. Bronchovascular bundle thickening and hilar lymphadenopathy were only seen in the lymphoid leukemias; nodules were also more common in these patients, although the difference was not statistically significant. Parenchymal nodules and masses are the most frequent finding in patients with leukemias in our study. This is consistent with work from Heyneman et al, who also found nodules to be the most common finding in their series.4 Of note, we saw 3 cases of centrilobular nodules, all in patients with lymphoid leukemias. While radiologists typically associate centrilobular nodules with infectious or inflammatory processes, these can rarely result from infiltration of leukemic cells along the bronchioles.8 Indeed, centrilobular nodules were found frequently in previous work focusing on lymphoid leukemias.7,9
We also found bronchovascular bundle thickening to be a unique feature in lymphoid leukemias and likely reflects the predilection of this malignancy for peribronchial lymphatics. This finding is consistent with prior work showing a high frequency of bronchovascular bundle thickening in lymphoid leukemias.8 Notably, the presence of septal thickening did not differ substantially between myeloid and lymphoid leukemias, although both presumably are related to perilymphatic infiltration. A substantial minority (43%) of patients in our series presented with clinical symptoms of infection (ie, fever and cough). With those symptoms in an immunocompromised patient, consolidations or centrilobular nodules are typically assumed to represent infections. However, recurrence or progression of the patient’s primary malignancy should always be kept in mind. In particular, in cases of patients with lymphoid leukemias presenting with centrilobular nodules, radiologists may suggest leukemic infiltration if the nodules increase concomitantly with worsening disease elsewhere and/or do not respond to antimicrobial therapy. In patients with any type of leukemia presenting with consolidations, radiologists should suggest this diagnosis if the abnormality persists or progresses despite appropriate antimicrobial therapy. The differential considerations for the latter scenario would also include aggressive atypical infections such as mucormycosis, and tissue sampling will likely be necessary to establish the diagnosis. It is important to distinguish infections from progression of leukemia, as these require substantially different management (ie, antifungals or antibiotics vs chemotherapy). Additionally, these 2 possibilities come with substantially different prognoses, as patients may recover from acute infections, while progression of
ARTICLE IN PRESS 4
B. Gosangi et al. / Current Problems in Diagnostic Radiology 00 (2019) 1 4
malignancy despite prior lines of chemotherapy portends refractory disease. Being a retrospective study, we are limited by inconsistent CT protocols and differing clinical management. Our sample size is small, although that is expected given the rarity of this entity. On the other hand, all cases in our series were proven pathologically to represent leukemic infiltration of the lungs. In conclusion, lung involvement by leukemia may manifest in a number of ways, ranging from focal lesions such as nodules, masses, and consolidations, to diffuse abnormalities such as septal and bronchovascular bundle thickening. Importantly, manifestations of consolidations and centrilobular nodules mimic infections, and in our series a substantial minority of patients presented with infectious symptoms. Radiologists should consider leukemic infiltration in cases where suspected infectious findings do not resolve with antibiotic or antifungal therapy.
References 1. Klatte EC, Yardley J, Smith EB, et al. The pulmonary manifestations and complications of leukemia. Am J Roentgenol Radium Ther Nucl Med 1963;89:598–609. 2. Maile CW, Moore AV, Ulreich S, et al. Chest radiographic—Pathologic correlation in adult leukemia patients. Invest Radiol 1983;18:495–9. 3. Barcos M, Lane W, Gomez GA, et al. An autopsy study of 1206 acute and chronic leukemias (1958 to 1982). Cancer 1987;60:827–37. 4. Heyneman LE, Johkoh T, Ward S, et al. Pulmonary leukemic infiltrates. Am J Roentgenol 2000;174:517–21. 5. Tanaka N, Matsumoto T, Miura G, et al. CT findings of leukemic pulmonary infiltration with pathologic correlation. Eur Radiol 2002;12:166–74. 6. Potenza L, Luppi M, Morselli M, et al. Leukaemic pulmonary infiltrates in adult acute myeloid leukaemia: A high-resolution computerized tomography study. Br J Haematol 2003;120:1058–61. 7. Okada F, Ando Y, Kondo Y, et al. Thoracic CT findings of adult T-cell leukemia or lymphoma. Am J Roentgenol 2004;182:761–7. 8. Trisolini R, Lazzari Agli L, Poletti V. Bronchiolocentric pulmonary involvement due to chronic lymphocytic leukemia. Haematologica 2000;85:1097. 9. Carmier D, Dartigeas C, Dartigeas C, et al. Serious bronchopulmonary involvement due to chronic lymphocytic leukaemia. Eur Respir Rev 2013;22:416–9.