HRCT is considered a non-invasive reference method for the diagnosis of ILD as it gives good parenchymal detail. [21,22]
Semi-QS systems are characterized by disease extent estimation and putting a grade in which the higher grades are coincided with the higher disease percentage, while in quantitative methods, the actual percentage of involvement of the lung is used. [23] In this current study, we use the semi-quantitative scoring method described by Goldin et al. [13], the method described in the Scleroderma Lung Study.
Semi-QA methods, in multiple studies, were compared to QA using computer-aid approaches. Computer-aid methods were well correlated with the visual scoring methods for the detection of lung fibrosis together with assessment of the extent of disease and there was no intrareader variation encountered in visual scoring methods. [6, 24-27](5, 6)
In this current study, the total lung score was 14.8±8.78 (mean±SD) where GGO was 7.20±5.79, the fibrosis was 5.93±4.91 and honeycombing was 1.70±2.76.
Our results were near the results of Zexuan et al who studied the scleroderma patients using the scoring system used by Goldin et al, they found that the total score was 14.35 ± 6.18 where GGO was 7.15 ± 3.94, PF 5.58 ± 2.91, and HC 1.36 ± 2.66. But they put also emphysema in the scoring that was 0.28 ± 1.05.[28]
Our results of spirometry pass more with restrictive type with FVC % was 64.4±9.20 (mean ± SD) and FEV1/FVC% was 82.1±6.03 and a significant negative correlation was found between the total lung score, GGO score, and F score with the FVC (r= -0.687, p=0.001; r= -0.558, p=0.001& r= -0.406, p=0.026 respectively) but the HC score was negatively correlated with FVC but didn’t reach the significant level. This coincided with the study of Mena-Vázquez et al that showed the characteristics and the progression of ILD in patients with ARD between the period of 2015 and 2020. The FVC of all patients (n=204) was 72±16.6 (mean ± SD) and at the end of the study was 68.2±16.2(mean ± SD).[29]
Multiple studies showed the correlation between the PFT and CT lung scoring system using different modalities.
In the study of Ohno et al; who evaluated the severity of connective tissue disease by a new machine learning-based software for CT texture analysis in comparison with the qualitatively thin CT assessed by 2 chest radiologists; there were statistically significant correlations between pulmonary functions and all the quantitative indexes except nodular lesions (using new CT texture analysis) and disease severity index (using qualitative thin CT analysis that involved the severity and extent scores). [30]
In the study of Marten et al, all patients with collagen vascular disease (CVD) (n=52) had findings of ILD in thin-section CT. The extent of ILD was 36.3 ± 27.2% detected by the readers (the extent of reticulation was 27.0 ± 23.3%, the extent of GGO = 9.2 ± 17.0%, the extent of the coarseness of a reticular pattern was 1.1 ± 0.6%). There were significant correlations between the average extent of ILD in CT and FVC and FEV1 (r = –0.559, P = 0.0002 and r = –0.379, P = 0.014 respectively). The reticulation extent correlated moderately with FVC (r = –0.436, P = 0.005), but there was no significant correlation between reticulation extent and FEV1. There was no significant correlation between the GGO, and coarseness extent with neither FVC nor FEV1.[31]
In the systematic review of Van Royan et al, 17 studies compared the quantitative CT (QCT) scoring with PFT. [25, 26, 31-45] The correlation between the QCT and FVC ranged from R2 =0.048[p < 0.0001] to R2=0.504 [p < 0.0001]. [6]
We used the same score as Goldin et al who studied the HRCT scan on 162 scleroderma patients. They found that the F and GGO scores were the most common abnormalities in symptomatic SSc patients, and the extent of F score had a statistically significant negative correlation with FVC (r= - 0.22), DLCO (r= - 0.44), and TLC (r= - 0.36). They suggested that pure GGO on CT might be reversible as it represents inflammation because its extent was correlated with acute inflammatory cells in bronchoalveolar fluid and not correlated with PFT (r= 0.28). [13]
Also, in multiple studies that used the semi-quantitative scoring systems, the total score on HRCT was significantly and negatively correlated with FVC%, TLC, and DLCO. [43, 46-49]
The study of Wangkaew et al included 31 patients with SSc and used the HRCT score that classified the parenchymal abnormalities into four items: ground-glass opacity, lung fibrosis, bronchiectasis, and honeycombing. In concordance with our study, they found significant inverse correlation between FVC % and total ground glass scores (r=−0.43; P<0.05), total Fibrosis score (r=−0.56; P<0.01), total bronchiectasis score (r=−0.43; P<0.05) and total HRCT scores (r=−0.52; P<0.01) together with significant correlation between HRCT scores and O2 saturation (r=−0.47; P<0.01) and ESR (r=0.38, P<0.05).[50]
Our study agreed with the study of Pandey et al in that the total HRCT score was correlated with elevated pulmonary arterial pressures (PAP). Their study showed a significant relationship between the peak pulmonary artery pressure and the total CT score (p<0.0001). They stated that the fibrotic score was the most predictive factor of PH on ECHO, which can help in the screening of patients with scleroderma for pulmonary hypertension (PH) by the extent of pulmonary fibrosis. In our study, the F score didn’t reach a significant level (p=0.059). They also stated that some patients had significant elevations in the PAP although they had small or no lung fibrosis. this could explain that the development of PH in scleroderma patients is multifactorial, e.g., pulmonary vasculopathy and capillary bed obliteration, and lung fibrosis could be the main factor. [51]
There was different variation in the results of EPASP in multiple studies that included different types of CVD. The study by Mukerjee et al found that the EPASP on echo (mmHg) was 39 ± 15 in scleroderma patients without lung fibrosis and 46 ± 18 in scleroderma patients with lung fibrosis detected by HRCT.[52] In the study of Mohammed et al, the mean EPASP in the studied SLE patients was 31 ± 5.1 mmHg with no significant difference between the diseased and control group. [53] These variations may be due to the different groups studied.
Also, multiple studies assessed the RV GLS in RA patients. Fine et al found that the RV GLS was reduced in RA patients in comparison with normal participants (−17.9 ± 4.7% versus −20.7 ± 2.4% and p < 0.001). [54]
Opposite to the study of Meune et al, which was done on 27 RA patients, there was no statistically significant difference between the control group and RA patients regarding systolic strain. This may be explained by using tissue Doppler and not STE like others. STE is a highly sensitive technique to measure the strain.[55]
In SSc, functional and structural abnormalities of the RV have been reported in about 21% of the patients.[56] Also, SSc patients with preserved EF had impaired GLS as compared with the controls (−18.2 ± 1.8% versus −21.3 ± 1.7%, respectively, p > 0.01), and the impaired GLS values correlated with the disturbance in the functional capacity and rhythm.[57]
Mukherjee et al hypothesized that before the development of obvious RV failure, there are subclinical changes in the regional RV strain that consider the precursor for impending myocardial dysfunction. They reported that the systolic RV longitudinal strain (speckle-derived) measures were impaired in scleroderma patients compared to controls (−17.7% vs. −20.4%, respectively).[58]
We found a lot of variability between the studies regarding disease scores and the variables included in the current study.
In one study that investigated the acute phase reactants (APR); they found that all APR results were significantly higher in RA patients versus the controls together with significant correlations between serum levels and DAS28 score and the serum CRP, in-between the ARP tests, considered the most useful for evaluation of the disease activity RA patients. [59]
Our results coincided with the results of Williams et al, that there was no correlation between the level of CRP and SLEDAI score.[60]
A lot of variables were associated with higher HAQ‐DI in SSc patients including FVC <70%, CRP ≥10 mg/l, ESR ≥20 mm/h, skin score ≥15, albumin <40 g/l, and fist closure ≥1.26 cm. Other variables did not associate with high HAQ‐DI e.g., disease duration, pulmonary artery pressure, sex, and hemoglobin and creatinine levels. [61]
In the study of stojan et al, they found that ESR elevations are strongly associated with disease activity in SLE patients during the same visit, as measured by SELENA revision of SLEDAI and this is not with our study.[62]
The results of the study of Ali et al, who studied the patients with SSc, found that there was no significant correlation between the EScSG score of disease activity and different clinical items such as EPASP and areas of lung fibrosis (p>0.05). Also, there was a significant correlation between FEV1/FVC and HAQ-DI (r=0.439; p=0.007) with a significant weak correlation between DLCO/ alveolar volume and HAQ-DI (r= – 0.347; p=0.045). The HAQ-DI was also significantly correlated with Rodnan skin scores. They mentioned that this was an unexpected finding as these factors were usually affected when there is an impairment of the functional status. Also, they stated that the significant correlation of FEV1/FVC cannot be clearly explained unless the increased FEV1/FVC demonstrates the more severe restrictive lung disease.[63]
Another study found a significant correlation between HRCT scores with HAQ-DI (r=0.37; P<0.05) and between FVC and HAQ-DI (r=−0.43; P<0.05) but in our study the correlation didn’t reach a significant level.[50]
In the study by Baron et al, HAQ-DI and the Short-Form 36 physical component summary score (SF-36 PCS) in SSc patients had a weak significant correlation with EPASP by ECHO although the independent contribution of EPASP to predict HAQ-DI or SF-36 PCS scores was found to be insignificant in a multiple linear regression analysis.[64]
In studies done on SLE patients, it was found no significant correlation between PAP and SLEDAI. [53, 65]
We can explain these variations by the small number of patients in each group that need to be more evaluated by increasing the studying participants.
In concordance with our study, one study assessed the correlations between SDAI and DAS28 scores in RA patients and values of both RV GLS. It showed a significant correlation between both scores and value of RV GLS i.e., the more the disease activity, the worse of the RV GLS value (r = -0.731; p = <0.001and r = -0.682; p= <0.001) for DAS28 and SDAI scores respectively. The value of RV GLS for active RA patients was worse significantly (less negative) compared to patients in remission and the control group (p = <0.001) and its value didn’t differ between the control group and patients in remission.[66] Also, in the study of Gullo et al on early RA patients, global longitudinal and circumferential strain (GLS and GCS) were impaired as compared to the control group (both p < 0.001), and DAS28 was significantly correlated to GLS.[67]
The GLS is considered a good method, even in the presence of normal EF, for detection of the reduction of ventricular systolic function.[68] Also in one study using cardiac magnetic resonance in RA patients; mid-wall fibrosis in the absence of ischemia is increased in association with the increased disease activity.[69]
In the study of Ntusi et al, it was reported that the impairment of myocardial strain in SSc patients was significantly correlated with the activity of the disease [70] and it was associated with increasing the cardiovascular risk. [71]
There were some limitations in our study e.g., the number of participants was small in each group of diseases, not all diseases of ARD are involved, and the scoring system used didn’t include other CT abnormalities e.g., emphysematous changes in ARD. We recommend further studies on a large number of patients with ARD with the comparison between different types of CT scoring systems that include more abnormalities recording. The correlation between different parameters of echocardiographic measures and the HRCT scoring system together with the scoring of disease activity needs more evaluation.