Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Feb 1.
Published in final edited form as: Clin Imaging. 2019 Nov 27;59(2):119–125. doi: 10.1016/j.clinimag.2019.11.004

Clinical CT Underestimation of the Percentage Volume Occupied by Cysts in Patients with Lymphangioleiomyomatosis

Thomas C Larsen a, Amir M Hasani a, Shirley F Rollison a, Tania R Machado a, Amanda M Jones a, Patricia Julien-Williams a, Marcus Y Chen a, Joel Moss a, Han Wen a,*
PMCID: PMC6941194  NIHMSID: NIHMS1544740  PMID: 31816538

Abstract

Purpose:

To evaluate the accuracy of cyst score measurements by standard high-resolution chest CT (HRCT) in patients with lymphangioleiomyomatosis (LAM), using a short z-length ultra-high resolution re-scan (UH re-scan) as the reference. In cystic lung diseases, cyst score is derived from CT scans and defined as the percentage of the total lung parenchymal volume occupied by cysts, a biomarker which measures the severity of the disease.

Methods:

In a prospective study of 73 LAM patients, each patient received the standard HRCT chest scan and a short z-length UH re-scan. Cyst scores were acquired from both scans using a standard FDA-approved scoring software on the CT scanner.

Results:

The limited UH re-scan resolved small cysts that were not resolved in the HRCT. The HRCT-derived cyst scores were on average 59.6% of the reference values from the UH re-scan (p = 4.7e-25). The amount of under-estimation by HRCT varied from patient to patient, with an inter-quartile range of 29.8% and standard deviation of 20.7%. The overall trend was more pronounced underestimation for patients with lower cyst scores. For patients whose reference cyst scores below 15 (n = 29), the HRCT cyst scores were 46.9 ± 21.6% of reference values (p = 7.4e-12), while for the rest of the patients (n = 44) the HRCT cyst scores were 68.0 ± 15.3% of reference values (p=1.2E-19). Reconstructing the HRCT images to the resolution of the UH re-scan further widened the spread of the discrepancy between HRCT and reference values due to increased image noise, and did not provide accurate cyst scores.

Conclusion:

Cyst scores derived from standard high-resolution helical chest CT significantly underestimates the percentage lung volume occupied by cysts. This inaccuracy needs to be taken into consideration when cyst score is used as part of the CT assessment of the patient’s condition.

Keywords: Lymphangioleiomyomatosis, cyst score, cyst volume, cyst segmentation, ultra-high resolution CT, short z-length

1. Introduction

In a clinical protocol to monitor and treat patients with lymphangioleiomyomatosis (LAM), regular x-ray computed-tomography (CT) scans have been used to monitor the progress of the patients[1-4]. In particular, the cyst score, a Federal Food and Drug Administration (FDA)-approved CT-based imaging marker, serves as a quantitative measure of the extent of the lungs affected by the disease[1,2,5,6]. LAM is a rare disease that occurs primarily in women of childbearing age. It results in the growth of abnormal smooth muscle-like LAM cells in the lungs. These LAM cells destroy lung parenchyma and generate thin walled cysts[3,7,8]. For patients with LAM, cyst score is a useful tool for initial assessment of the condition as well as to track disease progression[8-10]. The standard high-resolution CT chest scan has been the gold standard for cyst score measurements, which facilitated the introduction of other scan protocols that reduce radiation dose with advanced iterative image reconstruction [11-16].

The focus of this study is on the accuracy of the cyst score by standard HRCT, using as reference an ultra-high resolution short z-length re-scan that provides extra tissue resolution [17]. The accuracy of cyst scoring depends on image resolution to differentiate between cyst, parenchyma, and image noise. The ability of HRCT scans to resolve cysts and the resulting accuracy of cyst score were evaluated in a group of LAM patients. Limited or reduced Z length scans have been used previously in abdominal, neck and chest CTs in both adults and children as a way to reduce radiation dose [18-24]. The specific UH re-scan of this study has been optimized for best resolution in a systematic chest phantom study previously [17]. Typically, the UH re-scan deposits 40% the dose of a chest HRCT scan into a short z-length of 2 cm in axial scan mode, resulting in doubling the resolution at the same contrast-to-noise ratio. With such resolution, small cysts are clearly resolved and some cyst wall structures become visible.

2. Method

2.1. Study Population

The CT study was approved by the National Heart, Lung, and Blood Institute Institutional Review Board. All patients gave written informed consent. The protocol was applied to a total of 89 consecutive patients diagnosed with LAM. Of this group 16 were excluded in an initial screening of the CT images due to problems during the breath hold of the scans, which caused substantial motion blurring in either the standard HRCT (n=5) or UH re-scan (n=11). The final study population of 73 patients were all female. The mean age was 49.4 at the time of the scan with a range of 27 - 70. The mean pulmonary function test result of forced expiratory volume in 1 second (FEV1) was 81.3% of predicted values with a range of 42% - 119%. The mean test result of diffusion capacity (DLCO) was 68.0% of predicted values with a range of 23% –126%.

2.2. CT Scan Parameters

CT scans were performed on the Canon/Toshiba Aquilion One Genesis scanner. The HRCT helical scan used a tube setting of 120 kV/R700 mA and a rotation speed of 0.275s giving a total scan time of 3.6s to cover 360 mm with a helical pitch of 0.813. The short z-length UH re-scan was an axial scan (static patient bed) with a tube setting of 120 kV/350 mA, two rotations at a speed of 1.5s giving a total scan time of 3s, and covering 20 mm[17].

2.3. Scan Protocol

Each patient was scanned by the standard HRCT scan while lying prone. Then, using the reconstructed HRCT images as reference, the position of the short z-length UH re-scan was prescribed in situ (Fig.1). The UH re-scan was generally positioned at the level of the aortic arch, in order to be consistent among study subjects. In the exception of two patients where lung nodules of interests were found by the HRCT scan, the UH re-scan was positioned to specifically target the nodules. Both scans require breath-holding. The HRCT is a 3.6 second breath hold, and the UH re-scan is a 6 second breath hold with a 3 second settle-down period before the scan starts.

Figure 1.

Figure 1.

Open in a new tab

Screenshot demonstrating the difference between the two prescribed scans on a chest. The standard HRCT scan (a) covers 360 mm length while the short z-length UH re-scan (b) covers 20 mm. By focusing radiation into the short z-length, the UH re-scan achieves 2.0 times the resolution of the HRCT scan at the same contrast-to-noise ratio.

The effective dose of the HRCT scan was 0.93 mrem and that of the UH re-scan was 0.38 mrem, based on the average dose-length product reported by the scanner and published conversion factors[25].

2.4. Image reconstruction and Cyst Score Measurements

HRCT standard reconstructed images had a pixel size of 0.78 mm and slice thickness of 1 mm per the standard of care, while the UH re-scan had a pixel size of 0.39 mm and slice thickness/spacing of 0.5 mm.

The HRCT scans were also reconstructed to the same resolution of the UH re-scan, namely 0.39 mm pixel size and 0.5 mm slice spacing, to determine whether the difference between the standard and the UH scans could be accounted for by the reconstruction resolution.

The HRCT covered the entire chest in width and length. For this comparison study, a 20 mm length of the HRCT stack that exactly matched the UH re-scan coverage was selected for cyst score measurements to ensure that the same lung tissue was evaluated by both scans.

Measurements were performed with an FDA-approved cyst scoring software provided by the vendor with the CT scanner. The proprietary software provides automated segmentation of lung space and cysts based on a CT attenuation threshold combined with artificial intelligence algorithms. The vendor-recommended threshold for cyst segmentation is −940 Hounsfield units (HU). In this study the automated segmentation was monitored visually for each patient and each scan to ensure the correct separation of lung parenchyma, bronchi and cysts. Occasional adjustments to the threshold were made in the range of −920 to −950 HU when segmentation with the default threshold apparently failed.

3. Results

Paired images of standard HRCT scan and reconstruction resolution versus UH re-scan are shown in Fig.2. The UH re-scan resolved cysts which are not visible in the HRCT scan. Additionally, Fig.2b shows an image from a HRCT scan reconstructed to the high resolution of UH re-scans. When compared to the UH re-scan image in Fig. 2c, the lung parenchyma appears grainy indicating increased noise level in the image.

Figure 2.

Figure 2.

Open in a new tab

Examples of paired comparison of the HRCT and the UH re-scan images of the exactly matched regions of the chest. The images come from 4 patients. Panels (a, d, f) are from HRCT and (c, e, g) are from UH re-scan. The UH re-scan maintained the contrast-to-noise level while providing a higher-level of resolution, which revealed small cysts not seen in the HRCT. Panel (b) is an HRCT scan reconstructed to the same resolution as the UH re-scan, where the noise level or graininess is substantially elevated, which obscures the small cysts.

Automated cyst segmentation for the three types of images are illustrated in Fig.3. Some small cysts are missed in the image from HRCT scan with standard reconstruction resolution (Fig.3a and b), which are correctly segmented in the UH re-scan image (Fig.3c and d). In the HRCT scan reconstructed to the high resolution of the UH re-scan (Fig.3e and f), an increased graininess is seen. Errors in segmentation are seen where some random low-signal pixels in the parenchyma are included as cysts, and random high-signal pixels in cysts are considered parenchyma. The segmented cysts appear fragmented when compared to Fig.3d and 3b.

Figure 3.

Figure 3

Open in a new tab

Automated cyst segmentation for the three types of images. (a) and (b) are the images from the HRCT scan reconstructed at standard image resolution, showing the original image in (a), and segmented cysts as the green high-lighted areas in (b). (c) and (d) are images from the UH re-scan, showing the original and segmented images respectively. (e) and (f) are images from the HRCT scan reconstructed to the same resolution as the UH re-scan. Green areas in (f) are designated as cysts by the auto-segmentation n software. Comparison between (b) and (d) illustrates that some small cysts visible in (d) are missed in (b). In (f) some pixels in the parenchyma and in cysts are segmented incorrectly due to random fluctuations of the signal on the pixel scale (noise in the image).

Quantitative results of the cyst scores for the three types of images are shown in Figure 4. Figure 4a is the plot of the cyst scores from the standard HRCT scan at standard reconstruction resolution versus the reference UH re-scan for all patients. Cyst score is expressed as a value from 0 to 100, equivalent to 100*(cyst volume)/(parenchymal volume). The HRCT cyst scores were 59.6 ± 20.7% (mean ± standard deviation) of the reference cyst scores (p = 4.7e-25), with an inter-quartile range of 29.8%. A trend of more underestimation by HRCT at lower cyst scores can be seen. A piecewise analysis with a cyst score of 15 as a mathematical boundary showed that for patients with reference cyst scores below 15 (n = 29), the HRCT cyst scores were 46.9 ± 21.6% of reference values (p = 7.4e-12). For the rest of the patients (n = 44), the HRCT cyst scores were 68.0 ± 15.3% of reference values (p=1.2E-19).

Figure 4.

Figure 4.

Open in a new tab

The limited-length HRCT cyst score versus the reference UH re-scan cyst score for the exactly matched region of the chest, for all patients. (a) Cyst scores from the HRCT scan reconstructed to the standard resolution are 59.6 ± 20.7% (mean ± standard deviation) of the reference values, with an inter-quartile spread of 29.8%. For reference cyst scores below 15, the HRCT cyst scores are 46.9 ± 21.6% of reference values, and above 15 the HRCT cyst scores are 68.0 ± 15.3% of reference values. HRCT values were all below reference values. (b) Cyst scores from the HRCT scan reconstructed to the same resolution as the UH re-scan are 141.1 ± 74.8% of the reference values, with an inter-quartile spread of 52.3%. A trend of overestimation at low cyst scores and underestimation at high cyst scores is visible.

Figure 4b is the plot of the cyst scores from the standard HRCT scan versus the reference UH re-scan, both reconstructed to the high resolution of the UH re-scan. The HRCT high-resolution cyst score were 141.1 ± 74.8% (mean ± standard deviation) of the reference values (p = 9.8e-5), with an inter-quartile range of 52.3%. An overall trend of overestimation at low cyst scores and underestimation at high cyst scores can be seen for the HRCT high resolution images.

4. Discussion

Clinical chest HRCT scans are used to evaluate patients with LAM disease, in which the cyst score measurement is an integral part of the assessment of a patient’s condition. Historically, CT scoring of the severity of LAM initially used a subjective determination of the percentage of lung with abnormality, which was shown to correlate with pulmonary function tests such as forced expiratory volume in one second (FEV1) and diffusion capacity (DLCO % of predicted)[2,26]. Subsequently, the CT index of the percentage of the lung volume occupied by cysts, or cyst score, was introduced with automated segmentation of the lung space based on a threshold of CT attenuation[27]. This index removed the subjective factors in the evaluation. It is shown to correlate to various degrees with different types of pulmonary function tests in a number of clinical studies of LAM[2,6,10,27,28].

In this study, we found that the standard HRCT scan with standard reconstruction resolution underestimated the cyst score, due to inability to resolve small cysts, which became visible in high resolution images from a UH re-scan. The UH re-scan was optimized in a previous study on a chest phantom that contained lung mimicking and control foams [17]. That study concluded that the UH re-scan maintained the same noise level as the standard HRCT images but at twice the resolution. Ideally if the HRCT vs UH re-scan data (Fig.4a) converged to a narrow line, then the underestimation can be corrected systematically with a mathematical formula, without the need to perform the UH re-scan on individual patients. However, the data points were wide spread, with an inter-quartile range of 30%, indicating that the amount of underestimation varied from patient to patient and needs to be assessed on an individual basis. As the UH re-scan resolution is not microscopic, it is probable that there are still cysts that are below the resolution of the UH re-scan. So, the “real cyst score” may still be higher than the values from the UH re-scan, but not directly measurable by imaging until an in vivo microscopy technology becomes available.

With regard to the question whether the difference between the standard HRCT scan and the UH re-scan is simply due to the resolution of image reconstruction, this was assessed quantitatively in the previous chest phantom study[17], which found that the HRCT scan reconstructed at the resolution of the UH re-scan had 2.4 times the noise level of the UH re-scan. The increase in noise was consistent with the relative levels of dose concentration between the two scans. In this study, while the noise level could not be directly measured in the parenchyma due to the underlying tissue texture signal, an increased graininess can be seen in the HRCT scans reconstructed to the UH resolution. At this resolution for the HRCT scan, automated cyst segmentation had difficulty identifying noisy pixels as cysts or parenchyma. A wide spread of discrepancy was seen between the cyst scores from the HRCT scan and the UH re-scan reconstructed to the same resolution, with an inter-quartile range of 52% and a standard deviation of 75%. Therefore, raising the reconstruction resolution of the HRCT scan did not improve cyst score accuracy.

The UH re-scan protocol is within the hardware and software capability of current clinical scanners, and therefore does not require any hardware or software changes. It can be created readily on standard clinical scanners. The main limitation is that it only covers a fraction of the lung, and cannot be performed on the whole lung without incurring a large dose. Measurements from the UH re-scan cannot be directly correlated with lung function tests and other clinical tests that reflect the performance of the whole lung. The immediate value may lie in more accurately tracking cyst development over time. Current research is focused on developing appropriate image analysis methods to extrapolate the UH re-scan finding to the whole lung, in order to obtain measures that can be correlated with other clinical tests.

Highlights.

Lymphangioleiomyomatosis (LAM) is a rare disease that occurs primarily in women of childbearing age. It is characterized by the presence of thin walled cysts (cavities) in the lung parenchyma. The percentage volume of the parenchyma occupied by cysts, called the cyst score, is an essential radiological index for assessing the condition. Although clinical high-resolution CT chest scan has been the gold standard for cyst score measurements, we found through an ultra-high resolution short z-length CT study in a group of 73 patients, that the gold standard underestimates the cyst score by 40.4% (p = 4.7e-25), due to un-resolved small cysts. The amount of underestimation varied substantially from patient to patient (inter-quartile range was 29.8%). Simply reconstructing the standard high-resolution scan to ultra-high resolution did not improve the accuracy of the cyst scores. Data are summarized in the graphs:

Acknowledgements

This work is supported by the Division of Intramural Research, National Heart, Lung and Blood Institute, National Institutes of Health, USA.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • [1].Gopalakrishnan V, Yao J, Steagall WK, Avila NA, Taveira-DaSilva AM, Stylianou M, et al. Use of CT Scanning to Quantify Progression and Response to Treatment in LAM. Chest 2019. doi: 10.1016/j.chest.2019.01.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [2].Avila NA, Kelly JA, Dwyer AJ, Johnson DL, Jones EC, Moss J. Lymphangioleiomyomatosis: Correlation of Qualitative and Quantitative Thin-Section CT with Pulmonary Function Tests and Assessment of Dependence on Pleurodesis. Radiology 2002;223:189–97. doi:10/dz8wjn. [DOI] [PubMed] [Google Scholar]
  • [3].Johnson SR, Taveira-DaSilva AM, Moss J. Lymphangioleiomyomatosis. Clin Chest Med 2016;37:389–403. doi:10/gfv3sr. [DOI] [PubMed] [Google Scholar]
  • [4].Taveira-DaSilva AM, Steagall WK, Moss J. Lymphangioleiomyomatosis. Atlas Genet Cytogenet Oncol Haematol 2009;13:751–7. [PMC free article] [PubMed] [Google Scholar]
  • [5].Yao J, Taveira-DaSilva AM, Colby TV, Moss J. CT Grading of Lung Disease in Lymphangioleiomyomatosis. Am J Roentgenol 2012;199:787–93. doi: 10.2214/AJR.11.7888. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [6].Schmithorst VJ, Altes TA, Young LR, Franz DN, Bissler JJ, McCormack FX, et al. Automated Algorithm for Quantifying the Extent of Cystic Change on Volumetric Chest CT: Initial Results in Lymphangioleiomyomatosis. Am J Roentgenol 2009;192:1037–44. doi: 10.2214/AJR.07.3334. [DOI] [PubMed] [Google Scholar]
  • [7].Moir LM. Lymphangioleiomyomatosis: Current understanding and potential treatments. Pharmacol Ther 2016;158:114–24. doi:10/gd6b7s. [DOI] [PubMed] [Google Scholar]
  • [8].Harari S, Torre O, Cassandro R, Moss J. The changing face of a rare disease: lymphangioleiomyomatosis. Eur Respir J 2015;46:1471–85. doi: 10.1183/13993003.00412-2015. [DOI] [PubMed] [Google Scholar]
  • [9].Taveira-DaSilva AM, Stylianou MP, Hedin CJ, Hathaway O, Moss J. Decline in Lung Function in Patients With Lymphangioleiomyomatosis Treated With or Without Progesterone. Chest 2004;126:1867–74. doi: 10.1378/chest.126.6.1867. [DOI] [PubMed] [Google Scholar]
  • [10].Paciocco G, Uslenghi E, Bianchi A, Mazzarella G, Roviaro GC, Vecchi G, et al. Diffuse Cystic Lung Diseases: Correlation Between Radiologic and Functional Status. Chest 2004;125:135–42. doi: 10.1378/chest.125.1.135. [DOI] [PubMed] [Google Scholar]
  • [11].Nishio M, Koyama H, Ohno Y, Negi N, Seki S, Yoshikawa T, et al. Emphysema Quantification Using Ultralow-Dose CT With Iterative Reconstruction and Filtered Back Projection. Am J Roentgenol 2016;206:1184–92. doi: 10.2214/AJR.15.15684. [DOI] [PubMed] [Google Scholar]
  • [12].Nishio M, Matsumoto S, Seki S, Koyama H, Ohno Y, Fujisawa Y, et al. Emphysema quantification on low-dose CT using percentage of low-attenuation volume and size distribution of low-attenuation lung regions: Effects of adaptive iterative dose reduction using 3D processing. Eur J Radiol 2014;83:2268–76. doi: 10.1016/j.ejrad.2014.09.011. [DOI] [PubMed] [Google Scholar]
  • [13].Gierada DS, Pilgram TK, Whiting BR, Hong C, Bierhals AJ, Kim JH, et al. Comparison of Standard- and Low-Radiation-Dose CT for Quantification of Emphysema. Am J Roentgenol 2007;188:42–7. doi: 10.2214/AJR.05.1498. [DOI] [PubMed] [Google Scholar]
  • [14].Koyama H, Ohno Y, Yamazaki Y, Nogami M, Murase K, Onishi Y, et al. Quantitative and Qualitative Assessments of Lung Destruction and Pulmonary Functional Loss from Reduced-Dose Thin-Section CT in Pulmonary Emphysema Patients. Acad Radiol 2010;17:163–8. doi: 10.1016/j.acra.2009.08.009. [DOI] [PubMed] [Google Scholar]
  • [15].den Harder AM, de Boer E, Lagerweij SJ, Boomsma MF, Schilham AMR, Willemink MJ, et al. Emphysema quantification using chest CT: influence of radiation dose reduction and reconstruction technique. Eur Radiol Exp 2018;2:30. doi: 10.1186/s41747-018-0064-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [16].Hu-Wang E, Schuzer JL, Rollison S, Leifer ES, Steveson C, Gopalakrishnan V, et al. Chest CT Scan at Radiation Dose of a Posteroanterior and Lateral Chest Radiograph Series: A Proof of Principle in Lymphangioleiomyomatosis. Chest 2019;155:528–33. doi: 10.1016/j.chest.2018.09.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [17].Larsen TC, Gopalakrishnan V, Yao J, Nguyen CP, Chen MY, Moss J, et al. Optimization of a secondary VOI protocol for lung imaging in a clinical CT scanner. J Appl Clin Med Phys 2018;19:271–80. doi: 10.1002/acm2.12354. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [18].Weiss J, Maurer M, Ketelsen D, Notohamiprodjo M, Zinsser D, Wichmann JL, et al. Effect of reduced z-axis scan coverage on diagnostic performance and radiation dose of neck computed tomography in patients with suspected cervical abscess. PLoS ONE 2017;12. doi: 10.1371/journal.pone.0180671. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [19].Michalakis N, Keyzer C, De Maertelaer V, Tack D, Gevenois PA. Reduced z-axis coverage in multidetector-row CT pulmonary angiography decreases radiation dose and diagnostic accuracy of alternative diseases. Br J Radiol 2014;87. doi: 10.1259/bjr.20130546. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [20].Fefferman NR, Roche KJ, Pinkney LP, Ambrosino MM, Genieser NB. Suspected Appendicitis in Children:Focused CT Technique for Evaluation. Radiology 2001;220:691–5. doi: 10.1148/radiol.2203001826. [DOI] [PubMed] [Google Scholar]
  • [21].Rhea JT, Rao PM, Novelline RA, McCabe G. A focused appendiceal CT technique to reduce the cost of caring for patients with clinically suspected appendicitis. Am J Roentgenol 1997;169:113–8. doi: 10.2214/ajr.169.1.9207509. [DOI] [PubMed] [Google Scholar]
  • [22].Rao PM, Rhea JT, Novelline RA, McCabe CJ, Lawrason JN, Berger DL, et al. Helical CT technique for the diagnosis of appendicitis: prospective evaluation of a focused appendix CT examination. Radiology 1997;202:139–44. doi: 10.1148/radiology.202.1.8988203. [DOI] [PubMed] [Google Scholar]
  • [23].Kallen JA, Coughlin BF, O’Loughlin MT, Stein B. Reduced Z-axis coverage multidetector CT angiography for suspected acute pulmonary embolism could decrease dose and maintain diagnostic accuracy. Emerg Radiol 2009;17:31. doi: 10.1007/s10140-009-0818-6. [DOI] [PubMed] [Google Scholar]
  • [24].Parikh AK, Shah CC. Radiation Dose Reduction by Indication-Directed Focused z-Direction Coverage for Neck CT. Am J Neuroradiol 2016;37:985–9. doi: 10.3174/ajnr.A4672. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].Shrimpton PC, Jansen JTM, Harrison JD. Updated estimates of typical effective doses for common CT examinations in the UK following the 2011 national review. Br J Radiol 2016;89:20150346. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [26].Aberle DR, Hansell DM, Brown K, Tashkin DP. Lymphangiomyomatosis - CT, Chest Radiographic, and Functional Correlations. Radiology 1990;176:381–7. [DOI] [PubMed] [Google Scholar]
  • [27].Crausman RS, Lynch DA, Mortenson RL, King TE, Irvin CG, Hale VAE, et al. Quantitative CT predicts the severity of physiologic dysfunction in patients with lymphangioleiomyomatosis. Chest 1996;109:131–7. [DOI] [PubMed] [Google Scholar]
  • [28].Tobino K, Hirai T, Johkoh T, Fujimoto K, Kawaguchi A, Tomiyama N, et al. Difference of the progression of pulmonary cysts assessed by computed tomography among COPD, lymphangioleiomyomatosis, and Birt-Hogg-Dube syndrome. PLoSONE 2017;12. doi:10/gcn5pr. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES