Natural history of systemic sclerosis–related interstitial lung disease: How to identify a progressive fibrosing phenotype

Abstract

The natural history of interstitial lung disease in patients with systemic sclerosis is highly variable. Historical observational studies have demonstrated that the greatest decline in lung function in systemic sclerosis occurs early in the course of the disease; however, not all patients experience a decline in lung function even in the absence of treatment. Furthermore, among patients who do experience a decline in lung function, the rate of decline can be either rapid or slow. The most common clinical phenotypes of systemic sclerosis–related interstitial lung disease are, therefore, as follows: (1) rapid progressors, (2) gradual progressors, (3) stabilizers, and (4) improvers. This review summarizes the features of systemic sclerosis–related interstitial lung disease patients who are more likely to experience rapid progression of interstitial lung disease, as well as those who are more likely not to experience interstitial lung disease progression. Understanding the clinical, biological, and radiographic factors that consistently predict interstitial lung disease–related outcomes in systemic sclerosis is central to our ability to recognize those patients who are at heightened risk for interstitial lung disease progression. With new options available for treating patients with systemic sclerosis–related interstitial lung disease, it is more important than ever to accurately identify patients who may derive the most benefit from aggressive systemic sclerosis–related interstitial lung disease therapy. Early therapeutic intervention in patients with this progressive fibrosing phenotype may ultimately improve morbidity and mortality outcomes in patients with systemic sclerosis–related interstitial lung disease.

Introduction

Interstitial lung disease (ILD) occurs in the majority of patients with systemic sclerosis (SSc). In a recent systematic review of 50 publications reporting epidemiological data in SSc, ILD was present in 52% of the SSc patients in North America. ¹ The disease burden of SSc-ILD is substantial. ILD is not only the leading cause of death in SSc,^2,3 but it also adversely affects quality of life, diminishes functional capacity and can cause permanent disability. ⁴ However, an individual SSc patient’s experience with ILD varies considerably. In contrast with idiopathic pulmonary fibrosis (IPF), a disease in which patients almost universally experience an accelerated decline in lung function over time, patients with SSc-ILD have highly variable disease courses. Data from observational studies have demonstrated that while some patients with SSc-ILD experience a rapid decline in lung function, other patients experience a slow decline in lung function, and yet other patients experience no decline in lung function.^5,6

Data from randomized controlled trials (RCTs) for SSc-ILD have identified similar trends in disease course variation among patients.^7–9 In the Scleroderma Lung Study (SLS) I, which compared cyclophosphamide (CYC) with placebo for SSc-ILD, the majority of patients randomized to the placebo arm experienced a decline in their forced vital capacity % (FVC%) predicted over 12 months; however, approximately one-quarter of these patients actually had an improvement in their FVC%. ⁷ In the placebo arm of the SENSCIS trial, which compared Nintedanib with placebo for SSc-ILD, 34% of patients experienced an improvement in the FVC% after 12 months. ⁹ It is important to note that 49% of patients randomized to placebo in this study were receiving background mycophenolate mofetil (MMF) at the time of their enrollment. ⁹

Even among patients who receive treatment for SSc-ILD, the disease progresses at different rates. For example, in SLS II, which compared 24 months of MMF with 12 months of oral CYC followed by 12 months of placebo, 65% and 72% of participants randomized to the CYC and MMF arms, respectively, experienced an improvement in their FVC% predicted over the 2-year trial, while the remaining participants in both groups experienced worsening of their FVC% predicted. ⁸ The rate of improvement was highest in the first 12 months of the study for both treatment arms (Figure 1). ⁸

Figure 1.

The course of %-predicted FVC from 3 through 24 months by treatment arm in SLS II based on the joint model (N = 126). ⁸

The horizontal dotted line represents the average baseline FVC% predicted for both treatment arms based on the joint model. The vertical lines (dashed CYC; solid MMF) represent 95% confidence intervals. The * represents a significant change from baseline within treatment groups (p < 0.05).

Source: Reproduced with permission from Lancet Respiratory Medicine.

Taken together, these data from both observational studies and RCTs demonstrate that distinct ILD clinical phenotypes exist within the SSc disease state. This review article characterizes the predominant clinical phenotypes of SSc-ILD and focuses specifically on the progressive phenotype. This review also describes the clinical and biological factors that are associated with progression of ILD. Understanding the specific clinical and biological factors associated with distinct ILD phenotypes will improve our ability to stratify patients for appropriate therapy and to ultimately help personalize the care of all patients with SSc-ILD.

Defining disease progression in SSc-ILD

Paramount to defining and characterizing clinical phenotypes of SSc-ILD is reaching a consensus on how to define progression of ILD in SSc. While no valid definition of SSc-ILD disease progression exists, many experts believe that a reproducible decline in the measured FVC relative to baseline of ⩾10%, or a relative FVC decline of 5%–9% in association with a relative decline in diffusing capacity of carbon monoxide (DLCO) of ⩾15% represents progression. ¹⁰ However, changes in the radiographic extent of ILD on high-resolution computed tomography (HRCT) imaging of the chest are another way to define disease progression,^11,12 especially in circumstances where trends in pulmonary function testing (PFTs) are unclear. Moreover, external confounding factors are less likely to influence the results of HRCT imaging than of PFTs, where factors, such as the season, the time of day of testing, the patient’s overall wellness and energy levels, and the technician’s experience level, can individually affect PFT measurements.

In addition to changes in lung function and the radiographic extent of ILD, progression of ILD can also be defined in other ways, including the initiation of supplemental oxygen therapy for ILD, the need for lung transplantation, and death due to respiratory failure. Considering the aforementioned definitions of ILD progression, four distinct ILD phenotypes exist among patients with SSc (Table 1): (1) rapid progressors, (2) gradual progressors, (3) stabilizers, and (4) improvers. In clinical trials, a single primary endpoint is often used to define ILD progression; however, in clinical practice, all of the parameters described above and in Table 1 play an important role in defining disease progression in SSc-ILD and should be considered when making treatment decisions.

Table 1.

Distinct ILD clinical phenotypes exist in SSc.

	Rapid progressor	Gradual progressor	Stabilizer	Improver
FVC	Relative decline ⩾10%, or decline 5%–9% in association with ⩾15% decline within 1–2 years a	Relative decline ⩾10%, or decline 5%–9% in association with ⩾15% decline over >2 years a	Relative FVC decline <5%, or FVC increase <5%	Relative FVC improvement >5% b
HRCT	Increased extent of reticulations within 1–2 years a	Increased extent of reticulations over >2 years a	No change in the extent of reticulations	Decreased extent of reticulations
Supplemental oxygen c	Initiation within 1–2 years a	Initiation >2 years a	No need for supplemental oxygen or no increase in dose from the time of diagnosis	No need for supplemental oxygen or decreased dose from the time of diagnosis
Lung transplantation or death	Within 5 years a	>5 years a	No need for lung transplantation	No need for lung transplantation

ILD: interstitial lung disease; SSc: systemic sclerosis; FVC: forced vital capacity; HRCT: high-resolution computed tomography; SLS: Scleroderma Lung Study.

^aThese time periods are based on the time from the diagnosis of ILD; however, disease duration is often defined in different ways across studies (e.g. time of the SSc diagnosis from an SSc expert, the time from the onset of the first non-Raynaud’s symptom of SSc).

^bThe minimal clinically important differences (MCID) for improvement in FVC% based on SLS I and II data were 3.0%–5.3% at the cohort level. For an individual patient, however, using 5% as the threshold for improvement is a likely more conservative and reliable approach given the wide variation in FVC measurements.

^cInitiated for progression of ILD and not for other causes, such as progression of pulmonary hypertension (PH).

SSc-ILD rapid progressors

An important subgroup of patients with SSc-ILD will experience a rapid increase in the extent of reticulations (fibrosis) in their lungs, leading to a substantial loss of lung function early in the course of their disease. It is critical to identify “rapid progressors” early to initiate treatment promptly in an attempt to prevent the loss of lung function. Ideally, these patients should be flagged not only for early and aggressive treatment, but they should also be monitored more closely (at a minimum of every 3 months) for ILD progression and response to therapy.

Both observational studies and RCTs have identified specific factors associated with this worrisome clinical phenotype (Table 2). While each of these factors has been found to independently predict progression of ILD in SSc, having more than one of these factors could potentially augment the risk of progression.

Table 2.

Clinical, biological, and radiographic features associated with a rapid progressor SSc-ILD phenotype.

Demographic characteristics	Male sex, African American race, increased age
Disease-related features	Diffuse cutaneous disease, high MRSS at the time of ILD diagnosis; shorter disease duration
Pulmonary function tests	Moderate to severe restrictive physiology at the time of ILD diagnosis Decline in FVC and DLCO over 1–2 years
HRCT extent	Increased extent of reticulations at the time of ILD diagnosis
Serological profiles	Anti-Scl-70 antibody
Biomarker levels	High KL-6, high CCL-18, elevated CRP

SSc: systemic sclerosis; ILD: interstitial lung disease; MRSS: modified Rodnan skin score; FVC: forced vital capacity; DLCO: diffusing capacity of carbon monoxide; HRCT: high-resolution computed tomography; CRP: C-reactive protein.

Demographic factors

Males with SSc-ILD have poor prognoses and are more likely to experience ILD worsening early in the course of the disease.^13–15 Similarly, African American patients with SSc-ILD may also exhibit more severe ILD.^16–18 However, a recent analysis demonstrated that after adjustment for socioeconomic factors, African American race was not a significant risk factor for mortality in a relatively large cohort of SSc patients (N = 402) ¹⁸ Among SSc-ILD patients who participated in SLS I and II, there was no difference in the course of the FVC% predicted, nor long-term survival, between African American and non-African American participants, even after adjustment for baseline disease severity. ¹⁹ These findings suggest that confounding factors should be considered when interpreting the results of observational studies where patients have varying access to care, treatment, and follow-up.

In addition to sex and race, older age has been associated with increased mortality among SSc-ILD patients in both observational cohorts.^20,21 and in clinical trial cohorts. ²² Age-related comorbidities (e.g. cardiovascular disease, malignancy) may moderate the relationship between age and mortality in patients with SSc-ILD. Therefore, increased age alone may not be a defining feature of a “rapid progressor.” Nevertheless, older individuals require close monitoring for disease progression, given the higher mortality rate in this subgroup.

Disease-specific factors

In terms of disease-specific characteristics, patients with diffuse cutaneous sclerosis appear to have a higher risk of early SSc-ILD progression. ²¹ Similarly, a higher modified Rodnan skin score (MRSS) at baseline was associated with an increased risk of mortality in SSc-ILD patients from two clinical trial cohorts. ²² It is important to note that patients with limited cutaneous sclerosis have a nearly similar propensity for developing ILD as patients with diffuse cutaneous disease. ⁵ Furthermore, Clements et al. ²³ found there was no difference in the decline in lung function between limited and diffuse SSc-ILD patients randomized to placebo in SLS I after adjusting for baseline extent of radiographic fibrosis. Taken together, these findings suggest that although patients with diffuse cutaneous disease are more likely to experience rapid ILD progression early in the course of the disease, patients with limited cutaneous disease can also experience disease progression, especially when there is extensive radiographic fibrosis present at baseline.

Disease duration also affects the rate of SSc-ILD progression, and studies have demonstrated that patients with shorter disease duration have the greatest risk of ILD progression. ⁵ An observational study of 254 SSc patients with at least three serial FVC measurements found that patients with a disease duration of less than 2 years were more likely to have a faster decline in lung function compared with patients with a similar baseline FVC. ⁶ Shorter disease duration at the time of ILD diagnosis does not universally predict poor outcomes; ²⁴ however, it is unlikely for a patient with long disease duration (>5 years from ILD diagnosis) to experience rapid ILD progression at this disease stage.

Pulmonary function measurements

Baseline PFTs (e.g. FVC and DLCO) are often used to stratify patients with SSc-ILD. Patients with an FVC% predicted <50% during their first PFT had a 10-year survival of only 58%, compared with patients with an FVC% predicted >50% who had a 10-year survival of 87%. ⁶ Decreased FVC is associated with an increased risk of the development of clinically significant ILD, ²¹ as well as increased risk of mortality in SSc-ILD patients.^20,22,25 Similarly, a decreased DLCO is associated with an increased risk of the development of clinically significant ILD ²¹ and of death in SSc-ILD patients^20,22 It is important to note that patients who present with severe restrictive physiology at baseline have already lost a substantial proportion of lung function; therefore, their future rate of decline of FVC and DLCO may not be as rapid as other patients with less restriction at baseline. Patients with severe restriction at baseline can still be considered as “rapid progressors” because their ILD progressed rapidly prior to their initial presentation. In addition, this subgroup is at heightened risk for the development of other surrogate measures of ILD disease progression, including supplemental oxygen initiation, lung transplantation, and death.

In addition to the baseline FVC and DLCO, more recent studies have demonstrated that short-term trends in PFTs are important predictors of mortality in SSc-ILD patients.^20,22 Goh et al. ²⁰ discovered that a composite categorical decline (defined as either a decline in the FVC of ⩾10% or a decline in the FVC of 5%–9% in combination with a decline in the DLCO of ⩾15%) at 1 year predicted mortality in a cohort of SSc-ILD patients (N = 162) followed for a median of approximately 13 years (Figure 2). At 2 years (N = 140), changes in the DLCO appeared to have a greater impact on mortality than changes in the FVC for the whole cohort. ²⁰

Figure 2.

Increased long-term mortality among SSc-ILD patients who met the criteria for a compositive categorical decline (CCD) in lung function at 12 months (N = 162). ²⁰

The CCD was defined as either a decline in the forced vital capacity (FVC) of ⩾10% or a decline in the FVC of 5%–9% in combination with a decline in the diffusing capacity for carbon monoxide of ⩾15%.

Source: Reproduced with permission from Arthritis and Rheumatology.

Similarly, our group found that an increased decline in both the FVC and DLCO in the first 2 years after treatment initiation predicted an increased risk of mortality in two independent, clinical trial cohorts (N = 300). ²² In this study, the FVC and DLCO were measured every 3 months over 2 years, and these variables were entered into the survival model as time-dependent covariates. The trends in the FVC and DLCO were better predictors overall of long-term survival than the baseline FVC and DLCO measures. ²² We also explored the relationship between mortality and the composite categorical declines in the FVC and DLCO defined by Goh et al. ²⁰ However, these categorical declines did not consistently predict mortality in these two clinical trial cohorts, and this may have been due to the fact that only a paucity of patients who participated in SLS I and II experienced a DLCO decline of ⩾15% over 2 years. ²²

Radiographic features

Increased radiographic extent of ILD is also a feature of SSc patients with rapid ILD progression. In an earlier study, Goh et al. ²⁶ demonstrated that SSc-ILD patients with extensive disease on HRCT imaging (defined as >20% extent of ILD on HRCT, or as 10%–30% extent of ILD on HRCT in combination with an FVC <70%) had an increased risk of mortality. This staging system has been validated for its predictive significance in other cohorts. ²⁷ In a systematic review of 27 studies investigating variables that predict mortality and ILD progression in SSc, extent of disease on HRCT was the only variable found to independently predict both mortality and ILD progression. ²⁵

Increased radiographic extent of ILD may also be a feature of patients who are more likely to derive benefit from treatment for SSc-ILD. A post hoc analysis of SLS I data demonstrated that patients with greater than 50% or more reticular infiltrates in any lung zone on HRCT had an improved response to immunosuppression with CYC at 12 months. ²⁸

Although historical studies demonstrated that patients with a usual interstitial pneumonia (UIP) pattern of fibrosis on HRCT had worse outcomes, no substantial evidence exists to suggest that the pattern of fibrosis on HRCT or histology (e.g. non-specific interstitial pneumonia (NSIP) vs UIP) has a significant impact on disease progression or mortality in patients with SSc. ²⁹ However, transitions in ILD patterns may have prognostic importance in SSc patients. For instance, when our group evaluated changes in the quantitative extent of specific lung patterns related to ILD (e.g. fibrosis, ground glass, honeycombing) in SLS II participants, areas representing both lung fibrosis and ground glass transitioned to normal lung in patients treated with MMF and CYC, ³⁰ and overall improvements in the extent of ILD correlated with improvements in dyspnea and pulmonary function. ¹² These findings suggest that while increased extent of ILD on HRCT at baseline is a likely feature of a “rapid progressor,” it may also serve as a predictive marker associated with improved response to specific ILD therapies. ²⁸

Serological profiles

The presence of the anti-topoisomerase I antibody (anti-Scl-70) consistently predicts poor outcomes in patients with SSc-ILD. In a prospective, observational cohort of 266 patients with early SSc, the presence of the anti-Scl-70 antibody was the only baseline variable associated with an accelerated rate of decline in FVC within the first 3 years of follow-up. ³¹ In another cohort of 398 consecutive SSc patients followed for up to 15 years, the presence of the anti-Scl-70 antibody predicted the development of clinically significant ILD. ²¹ When interpreting the results of anti-Scl-70 testing, it is important to consider the method used to detect and measure these antibodies. Newer solid-phase multiplex approaches have high false negative rates. Indeed, a recent study compared anti-Scl-70 antibody testing in 129 patients undergoing three different testing methods: multiplex, enzyme-linked immunosorbent assay (ELISA), and immunodiffusion. ³² Among the 129 patients who tested positive for anti-Scl-70 by multiplex, only 51 of the patients were positive by ELISA, and 21 were positive by ELISA and immunodiffusion. ³² Therefore, it is critical to recognize the testing method employed when using anti-Scl-70 antibody testing to accurately identify patients with ILD who are more likely to experience rapid progression of SSc-ILD.

Detection of the anti-PM/Scl antibody is associated with the presence of ILD in SSc. ³³ Patients who harbor these antibodies typically present with signs and symptoms of myositis. SSc patients who possess anti-U11/U12 RNP antibodies often have severe ILD. ³³ Given the low prevalence of these antibodies and the infrequency of their evaluation in clinical practice, it is unclear whether the presence of these antibodies predicts rapid progression of ILD in SSc.

Candidate biomarkers

Prior studies have demonstrated that specific candidate biomarkers, such as interleukin (IL)-6, ³⁴ C-reactive protein (CRP), ³⁵ CCL-18,^36–38 CXCL4, ³⁹ and KL-6,^40,41 are associated with SSc-ILD progression in observational studies. The aforementioned studies did not adjust for treatment effect and used varying definitions of ILD disease progression. Furthermore, it is unclear when the biomarker was measured relative to treatment initiation (i.e. before treatment, during treatment, and after treatment). However, the collective results seem to suggest that high levels of these proteins in the peripheral circulation of a patient with SSc-ILD may predict future ILD worsening.

In clinical trial analyses, KL-6 levels were higher in SSc-ILD patients with evidence of active alveolitis in SLS I, ⁴² and in SLS II, higher KL-6 levels correlated significantly with increased extent of radiographic fibrosis and ILD on HRCT imaging, as well as with decreases in DLCO and total lung capacity (TLC). ⁴³ Moreover, patients with high KL-6 at baseline were more likely to experience ILD disease progression over the course of SLS II, despite treatment with immunosuppression. ⁴³ Similarly, patients with high CCL-18 at baseline were more likely to experience ILD disease progression in SLS II, despite treatment with immunosuppression. ⁴³ In addition, high CCL-18 levels were associated with an increased risk of mortality due to respiratory failure in patients randomized to the CYC arm, but not the MMF arm, of SLS II. ⁴³ These findings suggest that both KL-6 and CCL-18 are markers not only of disease severity but also of disease progression. Notably, both of these pneumoproteins decreased in response to treatment with immunosuppression in SLS II; therefore, the measurement of KL-6 and CCL-18 may have the greatest potential to inform prognosis if measured prior to treatment initiation and early in the course of the disease.

The integration of biomarker measurement into clinical practice may enable physicians to stratify SSc-ILD patients into clinical phenotype subgroups with greater precision than using clinical and demographic data alone. These biomarkers may also play a role in helping to predict response to specific therapies (e.g. immunosuppressive agents vs anti-fibrotics vs combination therapy), as well as helping to enrich clinical trial cohorts to evaluate the efficacy of novel SSc-ILD therapeutics.

SSc-ILD stabilizers and improvers

Even in the absence of therapy, certain patients with SSc-ILD exhibit stable lung function or can even have an improvement in lung function. As mentioned above, approximately one-quarter of patients randomized to the placebo arm of SLS I had an improvement in their FVC% predicted at 12 months. ⁷ If one defines ILD stability as an FVC increase or decrease <5%, well over half of the patients randomized to the placebo arm of SLS I met criteria for stability at 1 year. While it is presently difficult to predict which patients with SSc-ILD will exhibit ILD stability or improvement in the absence of treatment, the features of patients who are less likely to experience ILD progression are described further below (Table 3).

Table 3.

Clinical, biological, and radiographic features not typically associated with a rapid progressor phenotype.

Disease-related features	Limited skin disease; lower MRSS at presentation; Longer disease duration (>4–5 years from the diagnosis of ILD)
Pulmonary function tests	Stable pulmonary function over the first 4–5 years from the diagnosis of ILD
HRCT extent	None to moderate reticulations at baseline; stability in extent of reticulations over first 4–5 years from the diagnosis of ILD
Serological profiles	Anti-centromere antibody; anti-RNA polymerase III
Biomarker levels	High interferon chemokine score may predict improved response to immunosuppression in SSc-ILD 44

MRSS: modified Rodnan skin score; SSc: systemic sclerosis; ILD: interstitial lung disease; HRCT: high-resolution computed tomography.

Favorable disease-specific factors

Patients with limited cutaneous disease may be less likely to experience rapid ILD progression early in their disease course. Nihtyanova et al. ²¹ found that only 22% of limited cutaneous disease patients developed clinically significant ILD compared with 44% of diffuse cutaneous disease patients (p < 0.001). It is important to consider the duration of disease when using limited versus diffuse subtypes to predict the risk of ILD progression. For example, if a patient with diffuse SSc is very early in the course of their disease, they may have limited cutaneous disease on examination (prior to their development of diffuse disease). In addition, if a patient with diffuse SSc has a disease duration >4–5 years, they may also present with limited skin disease if their diffuse disease has already softened to a limited distribution. Careful history taking is of central importance in these scenarios to properly identify SSc patients with limited cutaneous disease. Patients with limited cutaneous sclerosis will generally experience progressive sclerosis of the hands, feet, face, and possibly forearms and lower legs. ⁴⁵ Their cutaneous sclerosis is typically not treatment responsive and remains stable over the course of their disease with only modest improvement over time.^23,46,47

Pulmonary function trends conferring positive ILD outcomes

It may be difficult to use PFT data alone to identify patients with a less progressive ILD phenotype early in the course of the disease. This is in part because when patients present with SSc, baseline PFTs prior to the disease onset are typically not available. A patient with an FVC% predicted of 85% on presentation is considered to have normal lung function; however, this patient may have had an FVC% predicted of 110% 3 years prior to their diagnosis. When interpreting lung function testing to stratify patients into SSc-ILD clinical subgroups, it may be more fruitful to consider the trends in pulmonary function. For instance, one study found that patients with a <5% change in their FVC over 4 years were less likely to experience ILD progression, which was defined as the initiation of supplemental oxygen, lung transplantation, or death. ²⁷ The same study found that patients who did not experience a decline in the FVC >10% or DLCO >15% at 1 year were also less likely to experience progression of ILD (negative predictive value of 92%). ²⁷ This study, combined with the recent works examining the relationship between PFT trends and mortality in SSc-ILD,^20,22 illuminates the importance of closely monitoring lung function in patients with SSc-ILD, especially within the first 5 years of the ILD diagnosis.

Advantageous radiographic features

The patients in the placebo arm of SLS I (N = 79) who experienced the slowest decline in lung function were those with none to moderate fibrosis on HRCT at baseline (mean annual decline in the FVC% predicted of 2.7%). ²⁴ Those with the greatest decline in FVC at 1 year had severe fibrosis on HRCT (mean annual decline in the FVC% predicted of 7.2%), and within this severe subgroup, the patients with the shortest disease duration (0–2 years) experienced the greatest decline in lung function. ²⁴ A single-center, observational cohort study found that patients with no evidence of fibrosis on HRCT at baseline were less likely to develop clinically significant ILD. ⁴⁸ When using HRCT imaging to determine the future risk of ILD progression, it is crucial to ensure that the HRCT is read by an experienced ILD expert, as early ILD can be missed or misinterpreted as atelectasis, especially when the imaging protocol is substandard (e.g. lack of prone imaging; thicker sections). In addition to baseline radiographic ILD extent, none to minimal progression of ILD on HRCT over the first 4–5 years from the time of the ILD diagnosis is also a defining feature of a less progressive ILD phenotype in SSc (Table 3).

Favorable serological profiles

A number of studies have demonstrated that patients who possess the anti-centromere antibody are less likely to exhibit a progressive SSc-ILD phenotype.^21,48,49 It is unclear whether the presence of anti-centromere antibody plays a direct role in protecting against progression of ILD or whether the presence of anti-centromere antibody is merely a surrogate marker for the presence of limited cutaneous disease. The latter seems more probable since the anti-centromere antibody rarely occurs in patients with diffuse cutaneous sclerosis, unless additional auto-antibodies associated with diffuse disease are present.

While anti-RNA polymerase III antibody is associated with diffuse cutaneous sclerosis, ³³ ILD may occur less often in patients who possess this antibody. ⁵⁰ No studies have adequately assessed the impact of anti-RNA polymerase III antibody positivity on progression of SSc-ILD.

Protective candidate biomarkers

Fewer studies have identified biomarkers that may protect against progressive pulmonary fibrosis in SSc. This may in part be due to publication bias, as journals may be more likely to publish “positive” biomarker studies. One could theorize that lower levels of the proteins mentioned above (e.g. CCL-18, KL-6) may be associated with a decrease risk of ILD; however, the exact threshold for defining high versus low protein level, nor the optimal timing for protein measurement, has not been established or validated across different SSc cohorts.

Our group recently discovered that patients with higher interferon (IFN) inducible serum protein levels were more responsive to immunosuppressive therapy with either MMF or CYC in SLS II. ⁴⁴ Specifically, patients with a high IFN chemokine score had an improved course of the FVC% predicted over 12 months in both SLS II treatment arms. ⁴⁴ In this study, the IFN chemokine score was based on a composite score of six different chemokines (MIG, IP-10, MCP-1, B2M, TNFR-2, and MIP-3b). ⁴⁴

In addition to measuring baseline protein markers, exploring how these proteins change in response to therapy could potentially serve as a meaningful way to stratify SSc-ILD patients in the future. In SLS II, we found that plasma CXCL4 levels decreased significantly from baseline to 12 months in all patients (CYC: p < 0.001; MMF: p = 0.006), and the patients with the largest decline in CXCL4 levels during the first 12 months of the study had an improved course of FVC% predicted from 12–24 months, even after adjusting for baseline disease severity and treatment arm assignment. ⁵¹ Measuring candidate biomarkers at earlier time points (after 3 months of therapy) might help assess treatment response prior to the detection of appreciable changes in lung function or radiographic fibrosis in patients with SSc-ILD.

Conclusion

Data from landmark observational studies ⁵ and more recent RCTs^7–9 illustrate that the natural history of ILD varies immensely in patients with SSc, regardless of treatment status. ⁵² No valid ILD progression prediction tools exist for this disease state; however, numerous studies have revealed important clinical, biological, and radiographic features that are associated with a more progressive fibrosing phenotype of SSc-ILD. The features that have consistently predicted poor outcomes among patients with SSc-ILD include older age, male sex, diffuse cutaneous sclerosis, high MRSS, more severe restriction on PFTs, increased decline in pulmonary function early in the disease course (<2 years), increased radiographic extent of ILD, and the presence of the anti-Scl-70 antibody. The features that may offset a risk for progression of ILD include limited cutaneous sclerosis, stability of pulmonary function on serial testing, none to mild radiographic fibrosis at baseline, and the presence of anti-centromere antibody. The discovery of novel biomarkers associated with ILD progression may expand our ability to further stratify patients into distinct clinical phenotypes.

While historical observational studies have enlightened our understanding of the most worrisome features of SSc-ILD patients, our ability to define SSc-ILD clinical phenotypes may evolve further as our understanding of the pathobiology of this disease state increases and as our capacity to offer diverse treatments to patients continues to grow. Factors considered to portend a poor prognosis in the past may become favorable features in patients treated with certain therapies. For example, even though most observational studies have demonstrated that increased extent of radiographic fibrosis at baseline is associated with worse SSc-ILD outcomes, patients with increased radiographic fibrosis at baseline in SLS I actually demonstrated the best treatment response to therapy with CYC compared to placebo.

The future of SSc-ILD therapeutic research depends on our ability to identify patients who are at the greatest risk for disease progression. Treatment effects are often blunted or entirely missed when patients who are unlikely to experience ILD progression are included in clinical trials. Cohort enrichment strategies in the past have failed because only a limited number of select variables were used to enrich the cohort ⁵³ when in reality a number of factors should be considered.

It is even more important in clinical practice to understand the clinical, biological, and radiographic features of patients who are at increased risk of SSc-ILD disease progression. These patients not only require early therapeutic intervention but they also need closer clinical monitoring and potential consideration for hematopoietic stem cell transplantation (HSCT) when more conventional therapies fail. Moreover, the patients themselves deserve to know their overall likelihood of experiencing disease progression. Promoting a dialogue with patients on their unique features that can affect their prognosis may alleviate the anxiety that often arises when a patient feels uncertain about his or her future.

In summary, recent advances in clinical and translational research have illuminated the most important factors that predict progression of ILD in patients with SSc. More work is needed to further refine our ability to stratify patients with SSc-ILD and identify those patients who need therapy at an early disease stage even prior to a decline in lung function. Future research is also needed to define distinct SSc-ILD subgroups who may demonstrate preferential response to specific ILD-targeted therapies. The findings of such research will allow us to personalize the care of patients with SSc-ILD and ultimately improve morbidity and mortality outcomes.