• Sarcoidosis
• Imaging/CT MRI etc

A major gap in the management of sarcoidosis is the lack of accessible and objective methods to measure disease activity.1 This has contributed to difficulties in deciding when treatment should be started or discontinued, particularly in fibrotic disease, and has hampered the evaluation of costly treatments such as biologics or cytokine modulators. The most commonly reported end point in clinical trials for pulmonary sarcoidosis is change in lung function measurements. However, this does not detect reduction in disease activity after treatment, which, in sarcoidosis, can also occur without change in lung function.

Ninety per cent of patients with sarcoidosis have pulmonary involvement, and high-resolution CT (HRCT) scan of the lung is performed in nearly all patients for diagnostic purposes. In addition, CT-pathological studies have shown that typical abnormalities like broncho-vascular nodularity, ground glass opacity and consolidation correlate to the presence of granuloma on open lung biopsies.2–4 Nodularity also associates strongly with bronchoalveolar lavage (BAL) cell count and serum-soluble IL-2 receptor (sIL-2R) levels,4 and profusion of interlobular septal thickening (IST) correlates with BAL lymphocytosis.5 ,6 We therefore explored if systematically measuring these HRCT abnormalities would allow quantification of disease activity in pulmonary sarcoidosis.

One hundred consecutive patients with pulmonary sarcoidosis from the Oxford Sarcoidosis Service who had HRCT scan performed from 2011 and fulfilling the diagnostic criteria proposed by WASOG were recruited. All patients were followed up for 1 year to record change in treatment and lung function. The study was approved by the South Central National Research Ethics Service Committee. Detailed study methods are found in online supplementary methods.

Thoracic HRCT scan and full pulmonary function testing were performed within 4 weeks of each other for all patients. Each HRCT was specifically scored for the presence, character and extent of sarcoidosis-related HRCT abnormalities (ground glass opacification (GGO), IST, nodularity, conglomeration, consolidation and intrathoracic lymphadenopathy) using predefined criteria from either the Fleischner Society glossary of terms or predefined criteria from previous peer-reviewed publications (see online supplementary table S1–S3). In an unselected proportion of patients—serum sIL-2R levels (n=21); serum ACE (n=72) and lung function at baseline and 12 months after CT scan (regardless of treatment) (n=40) were also performed. These parameters were chosen to reflect active macrophage and granuloma, T-cell activity and overall disease activity respectively. FVC change was measured as % change after a year from baseline, to represent disease reversibility and therefore, a surrogate of activity.

Scoring of the HRCT scans was performed by a fellowship-trained consultant thoracic radiologist (RB). To evaluate reproducibility of the final composite score, the first 85 scans were selected for scoring by a second radiologist based at a district general hospital (EH).

Ninety-seven patients had complete CT analyses. Demographic data and specific clinical characteristics are shown in table 1. One hundred per cent of patients had intrathoracic lymphadenopathy, 33% showed presence of GGO, 73% nodularity, 34% IST, 24% conglomeration and 20% consolidation (further details in online supplementary table S4). Thirty-four (35%) patients showed evidence of fibrosis. Four patients had evidence of pulmonary hypertension on CT scan, confirmed on right heart pressure assessment by echocardiography.

Extent scores for nodularity, IST, consolidation and GGO but not conglomeration or lymphadenopathy correlated with at least one of the surrogates of disease activity (see online supplementary table S5). Based on these findings, we excluded conglomeration and lymphadenopathy and took the sum of the scores of GGO, IST, consolidation and nodularity as the composite score for disease activity. We termed this the ‘CT Activity Score’ or ‘CTAS’. The CTAS score in our cohort ranged from 0 to 58, with a median of 10 (IQR 0–19) (see online supplementary figure S1A). Patients with fibrosis had higher CTAS compared with non-fibrotic ((median of 15.8 (IQR 8.5–22.0) vs 9.8 (4.0–15.0); p<0.001, Mann-Whitney U test) (see online supplementary figure S1B); and CTAS was higher with worse lung function (see online supplementary figure 1C–E). Examples of CT scans showing divergent CTAS are shown in online supplementary figure S2.

To provide a validation of the relationship of CTAS to disease activity, we examined if CTAS was able to predict the size of response to treatment. We reasoned that if CTAS reflected disease activity, then, the higher the CTAS value, the greater the ability of lung function (FVC) to improve with treatment. Twenty-nine patients went on to start or increase treatment specifically for their lung disease within 4 weeks of HRCT scanning. FVC values at 12 months (±8 weeks) after treatment were compared with pretreatment. We found that CTAS predicted the size of FVC response to treatment at 1 year (b=0.42 (CI 0.05 to 0.79); r2=0.17, p=0.03) (figure 1A). This was also the case for the fibrotic subgroup, where 20 of the 34 patients received treatment (b=0.32 (CI 0.0 to 0.6); r²=0.20, p=0.04) (figure 1B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Association between CT Activity Score (CTAS) and abbreviated CTAS (aCTAS) and changes in FVC at 12 months in patients who received new or increased treatment after CT scan, for all patients (A and C) and for the fibrotic subgroup (B and D). r and p Values shown are from linear regression calculations. Spearman correlation values are (A) r=0.40, p=0.03; (B) r=0.35, p=0.13; (C) r=0.42, p=0.02; and (D) r=0.34, p=0.13.

Comparing the CTAS with any one CT parameter (GGO, nodularity, consolidation and IST), we found only GGO and consolidation to predict treatment response (p=0.04 and 0.05, respectively), and at a weaker certainty than CTAS (p=0.03). Therefore, the composite measure of CTAS provides the strongest correlation with FVC response after treatment.

We also examined an abbreviated CTAS (aCTAS) comprising the sum of presence (1) or absence (0) of each of the four CT features (GGO, nodularity, IST and consolidation) and found that this produced similar correlation outcomes to CTAS for lung function and response to treatment (see figure 1C and D and online supplementary figure S3).

We found a strong correlation between radiologists for the CTAS (r=0.91; p<0.0001; Pearson's correlation test; n=85), and Bland–Altman analysis confirmed that there was no systematic deviation in measurements over the range of values (see online supplementary figure S4). There was no difference in aCTAS between the two radiologists (r=1.0; n=85).

We propose CTAS as a promising concept for the measurement of disease activity in pulmonary sarcoidosis. Our study was limited in size but is the first description of a numerical CT score for the extent of disease activity in pulmonary sarcoidosis. The validation component was limited to patients who were started on treatment during the year (n=29) and will benefit from further evaluation with a bigger cohort. Measurements of disease severity on HRCT (but not activity) have been previously published.7–10 CTAS is specifically designed to reflect the extent of disease activity regardless of the level of lung function impairment and fibrosis. Indeed, CTAS and aCTAS could be particularly useful in aiding decisions on whether to treat fibrotic sarcoidosis. A CT scan showing volume loss and established fibrosis has often been termed ‘burnt out’ fibrosis but may contain areas of active disease that may be driving the fibrotic process. We propose that a conclusion of ‘active’ or ‘inactive’ disease (with aCTAS), based on the absence or presence of any of these four CT abnormalities, could be a useful addition to an HRCT report for pulmonary sarcoidosis. A validation study will strengthen and possibly refine the score and provide an opportunity to bring this score into clinical practice.