The study was approved by the human research committee of the institutional review board and was compliant with HIPAA guidelines. The requirement for informed consent was waived. The inclusion criteria for the study were availability of images from combined PET/CT examinations for pulmonary lymphangitic carcinomatosis, confirmation of pulmonary lymphangitic carcinomatosis with follow-up chest CT or histopathologic examination, and histopathologic confirmation of the primary malignant disease. Patients who did not meet these criteria were excluded from the study.

The electronic radiology report database (Folio Views 4 software, Next Page) was searched for the key term “pulmonary lymphangitic carcinomatosis” among all chest PET/CT examinations. To improve the accuracy of the study, the search was limited to combined chest PET/CT. The search identified the cases of a total of 115 patients in the 4 years 2005–2008. Electronic medical records were reviewed to identify the cases of patients who had confirmatory histopathologic findings or had undergone follow-up CT or PET/CT to document unequivocal findings of pulmonary lymphangitic carcinomatosis, which were thickening of the bronchovascular bundle, interlobular septum, and subpleural interstitium. Of the 115 patients with suspected pulmonary lymphangitic carcinomatosis, 35 consecutively registered patients (15 men, 20 women; mean age, 64.5 years; range, 29–92 years) met the inclusion criteria. The primary site of malignancy and the results of histopathologic and pleural fluid analysis, if performed, were recorded.

All PET/CT examinations were performed with an integrated PET/CT scanner (Biograph Sensation 16, Siemens Healthcare), which has capability for 16-MDCT and PET with 3D lutetium oxyortho-silicate crystals. According to departmental protocol, patients were instructed to fast for 6 hours before scanning, and blood glucose was measured immediately before injection of an IV bolus of 15–20 mCi ¹⁸F-FDG. Static emission images were obtained 60 minutes after the injection.

Attenuation correction CT was initially performed with the following parameters: 40 mA, 120 kVp, 0.5 second per gantry rotation, detector configuration of 16 × 1.5 mm, and 10-mm slice thickness. PET images were acquired for 3–4 minutes per bed position for patients weighing less than 77 kg and for up to 7 minutes for patients weighing more than 104 kg. The attenuation correction CT images were fused with PET images, and fused PET/CT was used for visual interpretation and tumor volume measurements. A diagnostic neck, chest, and abdominal CT examination with IV administration of 120 mL iodinated contrast agent (iopamidol 370 mg 100 mL, Isovue, Bracco) also was performed. The imaging parameters for diagnostic CT examination were 120 kVp, 200 reference mAs, detector configuration of 16 × 1.5 mm, gantry rotation time of 0.5 second, 2.5-mm slice thickness at 2.5-mm slice interval, and pitch of 1.2.

Subjective analysis—The images were initially reviewed with the PACS of a diagnostic workstation (AGFA Impax ES, AGFA). The extent, distribution, and morphologic features of pulmonary lymphangitic carcinomatosis were reviewed on the CT images by two subspecialty-trained radiologists (8 and 10 years of experience). Pulmonary lymphangitic carcinomatosis was defined as thickening of axial, septal, and subpleural interstitium beyond the primary tumor in the lung, if a primary lung tumor was present. Changes in the lung beyond an obstructed bronchus were excluded.

Pulmonary lymphangitic carcinomatosis involving the lung parenchyma was called focal when it involved less than a lung lobe. Changes of pulmonary lymphangitic carcinomatosis affecting a whole lobe or the entire lung were labeled diffuse. The type (smooth or nodular) and distribution of interstitial thickening and the presence of lymphadenopathy, pleural effusion, and pleural nodules were recorded. A semiquantitative assessment of pleural effusion was made. An effusion was considered small if a small amount of fluid was present without or with only minimal relaxation atelectasis of the adjacent lung; moderate if it involved less than one third of the hemithorax, and large if fluid was present in more than one third of the hemithorax. Subjective analysis of the corresponding PET images was performed by the two radiologists to assess whether the area of pulmonary lymphangitic carcinomatosis exhibited increased FDG uptake.

Objective analysis—The images were transferred to an image-processing diagnostic workstation (Reveal-MVS, Mirada Solutions). The standardized uptake value (SUV) was calculated on this workstation on the basis of body weight and initial injected activity. A region of interest was manually marked on the CT image. The mean SUV was calculated on the workstation for the area of pulmonary lymphangitic carcinomatosis according to the formula SUV equals the mean region-of-interest activity divided by the injected dose. Mean SUV also was calculated for the corresponding segment of the contralateral, normal lung at the same slice level. Mediastinal blood pool activity was calculated at the level of the pulmonary artery. The standardized uptake ratios (SURs), normalized to the blood pool, between diseased lung and normal corresponding contralateral lung, normal lung and mediastinal blood pool, and lymphangitic lung and the blood pool were calculated. SUV depends on body weight and the anatomic site of the lung measured. To overcome this weakness, we measured the SUV in the contralateral lung at the site corresponding to the area of pulmonary lymphangitic carcinomatosis in the affected lung. We also calculated the SURs between blood pool and normal lung and between blood pool and diseased lung [8, 9].

The data were analyzed with Excel for Windows statistical software (Microsoft). Mean and SD for the various SUVs also were estimated with Excel software. The sensitivity for the two radiologists was calculated with the formula sensitivity = true-positive / (true-positive + false-negative). The specificity of PET/CT in the diagnosis of pulmonary lymphangitic carcinomatosis was calculated with the formula specificity = true-negative / (true-negative + false-positive), where the contralateral, normal lung was considered true-negative. Student's t test was used to determine statistical significance.

The study group consisted of 35 patients (15 men, 20 women; mean age, 64.5 years; range, 29–92 years). The primary malignant diseases were lung cancer (n = 29), prostrate cancer (n = 2), breast cancer (n = 1), pancreatic cancer (n = 1), small cell cancer of the head and neck (n = 1), and unknown primary tumor (n = 1). The presence of pulmonary lymphangitic carcinomatosis was confirmed histologically in two patients and on the basis of characteristic CT findings in 33 patients, all of whom underwent at least one additional sequential CT examination to confirm the findings. Histopathologically proven malignant pleural effusion was present in nine of the 35 patients.

Lung parenchyma—In 17 of the 35 patients (49%), pulmonary lymphangitic carcinomatosis was unilateral and diffuse, involving a whole lobe or one lung (Fig. 1A, 1B, 1C, 1D). In 13 patients (37%) pulmonary lymphangitic carcinomatosis was unilateral and focal, involving a part of a lobe (Fig. 2A, 2B, 2C, 2D), and in five patients (14%) it was bilateral and diffuse or focal. Bronchovascular bundle thickening and interlobular septal thickening were present in all of the patients. The interlobular septal thickening was smooth in five patients and nodular in 30 patients.

Both radiologists identified the area of uptake corresponding to pulmonary lymphangitic carcinomatosis on the PET/CT images of 30 of the 35 patients (sensitivity, 86%). Four of the five patients with uptake not detected with PET had focal (Fig. 3A, 3B, 3C, 3D) and one had diffuse pulmonary lymphangitic carcinomatosis of the right lung. All of these patients with pulmonary lymphangitic carcinomatosis missed on PET had a primary lung lesion (four patients) or metastatic nodules (one patient) close to pulmonary lymphangitic carcinomatosis, making distinction between the primary lesion and pulmonary lymphangitic carcinomatosis uptake difficult. The specificity of PET/CT was 100% because no abnormal uptake was identified in the contra lateral, normal lung or in the uninvolved portions of the ipsilateral lung.

Pleural effusion and nodules—Nine of the 35 patients had malignant pleural effusion. All of the effusions were moderate to large. Pleural nodules consistent with malignancy were present in four patients.

Lymphadenopathy—Hilar and mediastinal adenopathy was found on CT scans of 33 of the 35 patients. Increased FDG uptake was seen in enlarged hilar and mediastinal lymph nodes in 19 of the 33 patients.

The mean SUV in the area of pulmonary lymphangitic carcinomatosis was significantly greater than that in the contralateral normal lung (p < 0.0001). The mean SUV in the lung with pulmonary lymphangitic carcinomatosis was 1.37 ± 0.64 (range, 0.5–2.9), and that in the contralateral lung was 0.51 ± 0.29 (range, 0.1–0.8). The mediastinal blood pool activity measured in the pulmonary artery had a mean SUV of 1.49 ± 0.32 (range, 0.8–2.3). The SUR between the mediastinal blood pool and lymphangitic lung was significantly higher than the ratio between the mediastinal blood pool and normal contralateral lung (p < 0.0001) (Table 1). For the nine patients with histopathologically proven malignant pleural effusion, the mean SUV was 1.37 ± 0.43, and the ratio between the mediastinal blood pool and pleural effusion was 1.07 ± 0.20 (Table 1).