Background

Diffuse parenchymal lung diseases (DPLDs) affect the alveolar epithelium, pulmonary capillary endothelium, basement membrane, and perivascular and perilymphatic tissues. Compromising a heterogeneous group of diseases, it occurs when an abnormal healing response was induced by injury to the lungs [1, 2].

Investigations are aiming to reach the etiology of DPLD. The most important part of the workup of patients with DPLD is a detailed and proper history taking to identify the possible etiology including occupational or drug exposures, for signs of conditions like connective tissue and autoimmune diseases, sarcoidosis, and infection [3].

HRCT of the chest is the gold standard modality for diagnosis of DPLD. It provides 10 times more resolution than the conventional, as it reveals details that cannot otherwise be visualized [4]. The most common CT patterns of fibrotic interstitial lung disease (ILD) are ground-glass opacities, reticulations, traction bronchiectasis, and honeycombing. Those patterns have specific and broadly agreed definitions [5].

Pulmonary function tests usually show a restrictive defect in spirometry. DLCO-SB technique is used to assess the DPLD, as there is a thickening of the alveolar membrane and a diminished total lung capacity (TLC) due to interstitial processes [6].

Semi-quantitative scoring methods help to provide a more precise assessment of quantity and type of ILD abnormalities in HRCT chest. Using semi-quantitative scoring systems on HRCT has been shown consistently to be significantly and inversely correlated with forced vital capacity (FVC %), TLC, and DLCO predicted [7, 8]. One scoring system that has been used in several studies was developed and published by Warrick et al. [7].

In terms of assessing disease progression, it was found that patients with more severe fibrosis at baseline (or higher fibrosis scores) had a greater mean of decline in FVC and DLCO compared to groups with no or moderate fibrosis [9].

A clinico-radiological physiological scoring system is used to follow up patients with interstitial pulmonary fibrosis [10]. However, intending to reduce the frequency of performing HRCT during follow-up of patients with DPLD, we aimed to correlate between Warrick’s HRCT fibrosis score and DLCO-SB in DPLD, to assess the possibility of using DLCO alone as a tool for follow-up of DPLD patients to minimize the radiation exposure of the patients in HRCT chest (decrease need for radiological follow-up) or vice versa.

Methods

This prospective cross-sectional study recruited 89 patients having DPLD attending the Outpatient Clinics and Inpatient Departments of Pulmonology, Immunology, and Rheumatology at Ain Shams University Hospitals in the period from June 2019 to March 2020 as shown in Fig. 1.

Fig. 1
figure 1

Flow chart for patient selection

Full size image

Inclusion criteria

Patients older than 18 years old with DPLD either newly or previously diagnosed clinically dry dyspnea, dry cough and any other symptoms related to the etiology, and radiologically by HRCT Chest with different patterns without any of the exclusion criteria.

Exclusion criteria

Exclusion criteria were patients with comorbid cardiac, renal, or hepatic affection that may cause pulmonary venous congestion, patients with pulmonary or extrapulmonary malignancy, patients receiving radiotherapy and/or chemotherapy and/or biological therapy, patients with vanishing lung syndrome, and those refusing to participate in the study.

A written informed consent was taken from all patients prior to enrollment, and all patients were subjected to the following:

  1. 1-

    History taking for occupational, environmental exposures, autoimmune diseases, family history, and any relevant pulmonary or systemic symptoms.

  2. 2-

    Laboratory investigations: CBC, liver, kidney function, and collagen markers if not with the patient.

  3. 3-

    Echocardiography: To exclude heart as a cause of interstitial affection.

  4. 4-

    Spirometry and DLCO-SB: Performed according to the standard practice previously discussed in details [11, 12] using Viasys Healthcare Diffusion with built-in spirometer, D-97204 Hochberg, Germany. The spirometry indices recorded were forced expiratory volume in first second (FEV1), forced vital capacity (FVC), and FEV1/FVC ratio. Also, DLCO-SB % (best/predicted) was recorded after correction with the patient hemoglobin. By an online calculator available at https://www.uptodate.com/contents/calculator-diffusing-capacity-for-carbon-monoxide-dlco-correction-of-predicted-value-for-anemia-conventional-units.

  5. 5-

    HRCT chest: The study was done in the CT unit at Ain Shams University Hospitals by General Electric Bright Speed Elite 16 slices CT device, USA. Ethical approval was obtained to use the data stored on (PACS) picture archiving and communication system.

Method of HRCT examination

Thin sections were acquired with an interval of 1-2 cm between the two sets of images and were considered sufficient to detect abnormalities in diffuse lung diseases.

Fundamental technical protocols are as follows:

  • Slice thickness: 0.625-1.25 mm

  • Scan time: 0.5-1 s

  • kV: 120

  • mAs: 100-200

  • Collimation: 1.5-3 mm

  • Matrix size: 768 × 768 or the largest available

  • FOV: 35 cm

  • Reconstruction algorithm: high spatial frequency

  • Window: lung window (window level, −700 HU; window width, 1500 HU)

  • Patient position: supine (routinely), from lung apices to the bases.

  • Level of inspiration: full inspiration (routinely recommended).

Estimation of the total Warrick’s score 30 points maximum with extent score and severity score as previously described by Warrick et al. [7] (Table 1).

Table 1 Warrick’s total score and its subscores
Full size table

Total score was expressed as follows:

  1. a.

    Normal: zero point “no normal cases were involved in this work”

  2. b.

    Mild: less than 8 points

  3. c.

    Moderate: from 8 to 15 points.

  4. d.

    Severe: more than 15 points.

Also, two more scores were added

  1. a.

    Alveolitis score (ground-glass opacity (GGO) plus its severity score)

  2. b.

    Fibrosis score (CT abnormalities other than GGOsplus its severity score)

Ethical consideration

The study was approved by the Ethical Committee Board of Ain Shams University and was carried out in accordance with the Declaration of Helsinki.

Statistical data analysis

Data was analyzed using computer program SPSS (Statistical Package for the Social Science; SPSS Inc., Chicago, IL, USA) release 15 for Microsoft Window (2006). Qualitative variables were presented as percentage and quantitative variables were presented as mean ± SD. Student’s t test and Pearson’s correlation coefficient were used as the test of significance; P value less than 0.05 was considered as significant.

Results

This was a prospective study that recruited 89 patients presenting with DPLD, 74.2% of them were females; their mean age was 49.98 ± 12.56 years, range from 23 to 75 years old. Nine percent of the patients were farmers, and 15.7% were smokers. DPLD of unknown cause was found in 24.7% of patients (Tables 2) spiorometric and DLCO Parameters of the studied population were shown (Table 3).

Table 2 Demographics and associated risk factors or etiology of ILD among the studied population
Full size table
Table 3 Spirometry and diffusion parameters among the studied population
Full size table

As regards Warrick’s score, the total score mean ± SD was 17.38 ± 5.90 ranging from 7 to 29. Alveolitis score and fibrosis score were 2.99 ± 1.07 and 14.40 ± 5.32 respectively. Severity score and extent score were 8.63 ± 3.75 and 8.63 ± 3.08 respectively. GGO, irregular pleural margin, subpleural cyst, honeycombing, and septal and subpleural lines were found in 96.6%, 70.8%, 55.1%, 49.4%, 48.3% of the patients respectively (Table 4).

Table 4 Warrick’s score and its characteristics among the studied population
Full size table

Warrick’s score and its subscores (severity score, extent score, alveolitis score, and fibrosis score) were associated with a highly significant decrease in different pulmonary function indices (FVC, FEV1, TLC, and DLCO-SB). No significant correlation was found between FEV1/FVC ratio and Warrick’s scores (Table 5).

Table 5 Correlation between Warrick’s score and different pulmonary function indices among the studied population
Full size table

The current study showed a significant decrease in FVC in patients with honeycombing in HRCT compared to patients without honeycombing (59.33 ± 19.59 versus 68.51 ± 21.01) with P value 0.036. Also, a highly significant decrease in FVC was noted in patients with subpleural cyst than those without (57.4 ± 21.78 versus 72.02 ± 16.27) with P value 0.001 (Table 6). A highly significant decrease in FEV1 was noticed in patients with subpleural cyst than those without (55.69 ± 15.9 versus 69.57 ± 15.16 (with P value 0.001 (Table 7).

Table 6 Correlation between forced vital capacity and Warrick’s score indices
Full size table
Table 7 Correlation between forced expiratory volume in the first second and Warrick’s score indices
Full size table

This study showed a highly significant correlation between different DLCO-SB degrees of affection and total Warrick’s score degrees with P value 0.001. Most patients (86.8% from 38 patients) with severe degree of diffusion defect showed severe degree of total Warrick’s score (Table 8).

Table 8 Correlation between degree of diffusion lung capacity for carbon monoxide and degree of total Warrick’s score
Full size table

Patients with GGO in HRCT showed a significant decrease in DLCO-SB than patients without GGO in HRCT chest (43.3 ± 18.17 versus 68 ± 12.12) with P value 0.022. Also, a highly significant decrease in DLCO-SB was noted in patients with subpleural cyst and honeycombing with P value 0.001 and 0.002 respectively (Table 9).

Table 9 Correlation between DLCO-SB and Warrick’s score indices
Full size table

The validity of DLCO-SB for the prediction of GGO was assessed, and the area under the receiver operating characteristic (ROC) curve was 0.870. A value of ≤ 54.2 diffusion defect was statistically proposed as a cutoff value to predict presence of GGO (Fig. 2). Also, regarding the validity of DLCO-SB for the prediction of honeycombing in HRCT chest, the area under ROC curve was 0.685. A value of ≤ 43 diffusion defects was statistically proposed as a cutoff value to predict the presence of honeycombing (Fig. 3).

Fig. 2
figure 2

ROC curve and coordinates for the best cutoff DLCO-SB value to predict ground-glass opacity. ROC, receiver operating characteristic; AUC, area under the curve

Full size image
Fig. 3
figure 3

ROC curve and coordinates for the best cutoff DLCO-SB value to predict honeycombing. ROC, receiver operating characteristic; AUC, area under the curve

Full size image

Discussion

HRCT of the chest is the gold standard modality for the diagnosis of DPLD [4]. The most common CT patterns of fibrotic ILD are GGO, reticulations, tractions bronchiectasis, and honeycombing [5]. Pulmonary function tests usually show a restrictive defect in spirometry. As there is thickening of the alveolar membrane, DLCO-SB is also diminished [6].

This study aimed to evaluate the extent of DPLD using a semi-quantitative Warrick’s HRCT score to determine the disease severity and its possible correlation with PFT including spirometry and DLCO-SB. They have been the only parameters considered to date to follow up cases with DPLD aiming to replace one of them by the other either to decrease cost or decrease radiation exposure.

This prospective study recruited 89 patients diagnosed with DPLD, 74.2% of them were females; their mean age was 49.98 ± 12.56 years. The age group was matched with the results of Abou Youssuf H. et al.; however, the official ATS/ERS/JRS/ALAT guidelines 2011 stated that IPF occurs in the sixth and seventh decades [13, 14]. The higher incidence of DPLD in the female population in the studied group was consistent with Abou Youssuf et al. and Rifaat et al. who found out that the majority of the studied population were females [13, 15]. Additionally, it was in agreement with Hassan who studied the comparison of FVC, FVC/DLCO, and TLC/DLCO as an indicator for interstitial lung disease in patients with scleroderma [16].

DPLD was idiopathic in 24.7% of the cases. The observations of Kundu S et al. were similar; they concluded that unknown etiology was the most common cause of DPLD representing 38.04% [17]. This was in contrast with A Kheliouen. et al. who found that sarcoidosis was the most common disease, followed by ILD associated with connective tissue diseases. The discrepancy might be due to lack of open lung biopsy in the study population [18].

DPLD as one of the restrictive lung diseases affecting the alveolar-capillary membrane is associated with lower FVC% and DLCO-SB especially in severe degrees as shown by this work and others [13, 19]. A recent study was not totally in agreement with this work results in this issue, which might be related to the difference in the number of patients recruited and the variability in the characteristics of the studied populations [20].

FEV1, FEV1/FVC, and the TLC changes were consistent with Riad N et al. during their assessment of ILD with DLCO-SB and spirometry [21].

In the current study, Warrick’s score and its subscores (severity score, extent score, alveolitis score, and fibrosis score) were associated with a highly significant decrease in different pulmonary function indices (FVC, FEV1, TLC, and DLCO-SB). Previous studies showed the same results considering fibrosis score as interstitial score [22, 23].

Similar to the previous published data [24,25,26,27], we demonstrated that Warrick’s semi-quantitative score was able to determine the severity of DPLD correlated well with FVC, TLC, and DLCO.

In other words, the current study aimed to investigate the comparative value of DLCO-SB and HRCT in diagnosing DPLD, and showed a highly significant correlation between different DLCO-SB degrees of diffusion defects and total Warrick’s score degrees. These findings are consistent with Khanna et al. who found that most of the patients with severe fibrosis in HRCT were with severe diffusion defect [9].

Decline in spirometric indices and DLCO-SB were associated significantly with the different radiological patterns of CT chest. FVC was significantly decreased in patients with honeycombing and subpleural cyst in HRCT. Highly significant decrease in FEV1 was noted in patients with subpleural cyst. Also, this work demonstrated a highly significant decrease in DLCO-SB in patients with subpleural cyst and honeycombing and a significant decrease in DLCO-SB in patients with GGO.

These data are in accordance with Hussein et al. who found that FVC and DLCO-SB were significantly lower in fibrotic DPLD [28]. Also, Araki et al. who studied the effect of pulmonary cysts on DLCO-SB supported this work result [29]. Many other related studies went hand in hand with the current study results [30,31,32].

These significant correlations encouraged us to determine the cutoff values between DLCO-SB and the two extremes of HRCT structural changes of the lung, GGO, and honeycombing, to predict the value of DLCO-SB associated with these changes. A value of ≤ 54.2 diffusion defect was statistically proposed as a cutoff value to predict the presence of ground-glass opacity “alveolitis score” with better area under the ROC curve 0.870. Also, the validity of DLCO-SB for prediction of honeycombing in HRCT chest was assessed, the area under ROC curve was 0.685. A value of ≤ 43 diffusion defect was statistically proposed as a cutoff value to predict presence of honeycombing. So first baseline HRCT in patients with DPLD to determine the pattern and extent of radiological affection could be enough then follow up with DLCO-SB and less frequently with HRCT to limit radiation exposure and decrease the cost.

This study had some limitations; the small number of studied population and the shorter duration of the work with no follow-up data. HRCT other scores may be needed to correlate with the functional assessment of the lung on a larger number of patients. Longer follow-up periods are recommended to validate this study’s results and estimate the percentage of change of DLCO-SB with each HRCT radiological pattern and with each type of ILD.

Conclusion

Structural changes in the lung can be detected by HRCT chest, and functional changes can be assessed by spirometry and more precisely by DLCO-SB; in this work, both were significantly correlated. Since DPLD patients need lifelong follow-up, assessment of function and structure of the lung by HRCT chest and DLCO-SB at the first visit, with subsequent follow-up of lung function mainly by DLCO-SB to decrease follow-up times with HRCT appeared to be satisfactory with less cost and minimizing repeated radiation exposure.