The idiopathic inflammatory myopathies (IIM) are a group of acquired, heterogeneous, connective tissue diseases that mainly affect skeletal muscle. Characteristic clinical and histopathologic features allow their classification into polymyositis, dermatomyositis, and inclusion body myositis. The association between cancer and IIM has been reported in adults, although the pathogenesis remains elusive. Although cancer-associated myositis (CAM) has a worse prognosis in afflicted patients, its timely recognition is therapeutically important and contributes to morbidity and mortality. This article discusses several aspects of CAM including epidemiology, risk factors, autoantibody associations, pathophysiology, and screening.
The idiopathic inflammatory myopathies (IIMs) are a group of acquired, heterogeneous, connective tissue diseases that mainly affect skeletal muscle. Characteristic clinical and histopathologic features allow their classification into polymyositis (PM), dermatomyositis (DM), and inclusion body myositis (IBM). The association between cancer and IIM has been extensively reported in adults, although the pathogenesis remains elusive. Several epidemiologic studies have confirmed the increased risk of cancer in patients with myositis, particularly in DM patients compared with other myositis subsets. Although cancer-associated myositis (CAM) has a worse prognosis in afflicted patients, its timely recognition is therapeutically important and certainly contributes to morbidity and mortality. In this article, the authors discuss several aspects of CAM including epidemiology, risk factors, autoantibody associations, pathophysiology, and screening.
Epidemiology
Several epidemiologic studies have reported the association between cancer and myositis for nearly 100 years. Since 1916, when the first case reports of malignancies occurring in patients with both DM and PM were described, many studies have substantiated this association and attempted to understand the complex relationship between these 2 entities. Early reports included case–control studies and were limited by a lack of standardized definitions, referral bias, small sample sizes, and other methodological shortcomings. Moreover, amyopathic DM and necrotizing myopathy, 2 more recently described myositis subsets, were not well recognized. In the last decade, large population-based epidemiologic studies from the Scandinavian countries, Australia, and Scotland have confirmed the increased risk of malignancy in patients with inflammatory myopathy. Most studies estimate a frequency of cancer in adult patients with myositis between 10% and 30%, with DM representing the greatest risk factor.
In 2001, Hill described the temporal link between myositis and malignancy in a pooled analysis of national cancer registry databases using discharge diagnoses from Sweden, Finland, and Denmark. Standardized incidence ratios (SIRs), the observed incidence of malignancy in the myositis cohort compared with the expected rates in the general population, were 3.0 for DM and 1.4 for PM. Specifically 618 DM and 914 PM patients met Bohan and Peter criteria, with a cancer frequency of approximately 30% in DM and 15% in PM patients. Sixty percent of tumors were diagnosed after the diagnosis of myositis, and most cancers were detected within 1 year of patients developing myositis. As seen in earlier observational studies, the most common cancer types were adenocarcinomas, which accounted for 70% of all associated tumors. A major limitation of this study was the lack of biopsy-proven myositis and little ethnic diversity, suggesting that their results are mainly relevant for Caucasians. Another large epidemiologic study using state cancer registry databases and the national death index identified 74 (14%) cancer cases in a population-based retrospective cohort of 537 patients with biopsy-proven IIM. Malignancy was diagnosed concurrently (within 7 days) or after the diagnosis of myositis, and the overall risk for cancer in IIM was overall 2 to 3-fold higher (SIR 2.6), greatest in DM (SIR, 6.2 [confidence interval, CI, 3.9–10.0]) but also increased in PM (SIR, 2.0 [CI, 1.4–2.7]). In a metaanalysis, Zantos analyzed 1078 patients with PM/DM (565/513) from four different studies with a comparable number of controls. One hundred fifty-three myositis patients had cancer within 10 years of diagnosis, with the findings supporting an association of both PM and DM with cancer. An overall combined relative risk of cancer was 4.4 (9.5% CI 3.0, 6.6) and 2.1 (95% CI 1.4, 3.3) in DM and PM, respectively.
In contrast with large studies demonstrating a significant relationship between PM and cancer, Antiochos reported no association of cancer with PM. Similarly, Airio also observed a higher relative risk of cancer (SIR 6.5, 95% CI 3.9–10) among patients with DM but not in patients with PM in a population-based study of myositis in Finland.
Association Between Histologic Types of Cancer and Myositis
In DM, all histologic types of cancer have an increased risk, but adenocarcinomas are the most common. In contrast, hematological malignancies have the highest risk in PM, with no increase observed in the risk of squamous cell cancers and adenocarcinomas. Although a wide spectrum of cancer types can be associated with myositis, certain types are more frequent and may help guide the workup for occult malignancy. Most studies show the highest risk of lung, ovarian, cervical, colorectal, gastric, breast, and pancreatic cancer along with lymphomas in DM. In PM, the greatest increased risks were for lymphoma in most studies, and lung and bladder cancers in some studies, and there was no increased risk of ovarian, colorectal, stomach, or pancreatic cancers.
Temporal Relationship Between Time of IIM Diagnosis and Cancer Risk
Cancer can be diagnosed before, concurrent with, or after the diagnosis of inflammatory myopathy. The peak incidence of malignancy (PM/DM combined) is highest within the first year of myositis diagnosis (60%–70% cases) and decreases substantially each year thereafter. The overall interval with highest probability for tumor recognition is generally between 2 years preceding and 3 years following the diagnosis of myositis. In most studies, the risk of cancer in DM extends up to 5 years, while the cancer risk in PM appears to be shorter. In Hill’s study, the risk of cancer in PM was much lower after the first year of diagnosis, and fell to the expected baseline risk of cancer in the general population within 5 years of diagnosis. In DM, the higher risk of most cancers extended up to 5 years, and in colorectal and pancreatic cancer, the risk extended beyond 5 years. The risk of non-Hodgkin lymphoma in both DM and PM was only increased in the first year. A similar decreasing trend in the risk of malignancy was observed each year after the diagnosis of myositis (PM and DM) by Buchbinder, with SIR of 4.4 (CI 2.7–7.1) in the first year; 3.4 (CI, 2.3–5.1) between 1 and 3 years; 2.2 (CI, 1.3–3.9) between 3 and 5 years; and 1.6 (CI, 1.0 –2.6) beyond 5 years. This risk of cancer in DM reaches the baseline general population risk after 5 years when adjusted for age, gender, and calendar year. Of cancers diagnosed before DM, the risk was increased up to 2 years before diagnosis, whereas in PM there appears to be no increase risk before 1 year before diagnosis in most studies. One meta-analysis noted no increase risk of cancer before the diagnosis of PM.
Relationship of Age, Gender and Ethnicity with CAM
The mean age of CAM approximates 50 to 60 years of age, although the range is wide. The risk of cancer associated with DM appears higher (more than 3-fold in most studies) in patients 45 years and older, as compared with younger patients (2-fold increase for 15–44 years). In PM, older patients are more at risk in most studies, but Stockton showed the risk was higher in the 15 to 44 age group in PM as compared with patients 45 to 74 years old. Many studies show an increased risk of malignancy with increased age.
The risk of cancer is increased in both men and women, although many but not all studies suggest a slightly higher risk for men. The SIRs for men and women with DM were 3.3 (95% CI 2.5–4.4) and 2.8 (2.2–3.6), respectively. The SIR for PM was 1.4 (95% CI 1.0–1.8) for men and 1.2 (0.9–1.6) for women. There was no difference in the rate of sex-neutral cancers between male or female.
The epidemiology of CAM is very different in some nonwhite ethnicities. Breast, stomach, and nasopharyngeal cancer (NPC) are more common with DM in Korea, while NPC is the most CAM malignancy in Singapore, Hong Kong, southeastern China, and Taiwan.
Risk of Cancer in Other Myositis Subsets
One population based study found an increased risk for malignancy in IBM (SIR, 2.4 [CI, 1.2–4.9]), myositis associated with connective tissue disease (SIR, 4.6 [CI, 1.2–11.7]), and childhood myositis (SIR, 29.0 [CI, 3.5–105.0]). Others have also noted this malignancy association with juvenile DM (JDM), amyopathic DM (ADM), and IBM, with further substantiation in ADM. However, these associations need to be confirmed by larger population-based studies.
Survival in CAM
The type and stage of the underlying malignancy and the presence of severe myositis features such as respiratory insufficiency and dysphagia determine the prognosis of CAM. One-year survival and 5-year survival were significantly higher in non-CAM (95% and 92%) as compared with patients with CAM (88% and 66%). Increased mortality was also noted in an earlier study, with death rates of 75% in CAM (both PM and DM) compared with 12.5% in non-CAM ( P <.001), and death always resulted from progression of the malignancy. In patients where cancer and DM are simultaneously diagnosed, the disease course is more severe than those where the tumor is diagnosed before myositis.
Controversies Regarding the Relationship Between Cancer and Myositis
Although epidemiologic studies have confirmed the association between cancer and myositis, the question of increased cancer surveillance contributing to this relationship persists. This is doubtful for several reasons:
In DM the increased risk of cancer persists for several years of follow-up.
There is a clustering of cancer cases before the diagnosis of myositis.
Mortality due to cancer is increased in DM, which would not be affected by the fortuitous diagnosis of malignant disease.
An increased risk of malignancy in the time following the diagnosis of DM has been noted even when the initial year of follow-up is excluded.
In contrast, increased surveillance for cancer as well as misclassification of DM as PM could partially contribute to the greater risk in PM. A meta-analysis of case-control and cohort studies of myositis and malignant disease showed no increase in cancer detection before the diagnosis of PM, despite a raised risk after diagnosis. Similarly, only Caucasian patients with DM and not PM from northern New England showed a high risk of developing cancer. However, Buchbinder, who classified myositis histologically and was less likely to have misclassified DM and PM, showed a small but significant increased risk of malignant disease in PM. A slight but increased risk of malignant disease in PM seems likely.
Effects of CAM Treatment on Myositis
Although treatment of the tumor (surgery, chemotherapy, radiotherapy) may lead to remission of the myositis, many patients require long-term immunosuppression to control myositis even with cancer remission. Myositis can recur upon relapse of malignancy even years after initial presentation.
Risk factors for CAM
Identifying IIM patients at risk of developing malignancy has clear importance to the managing clinician. Older age, male sex, refractory or recurrent myositis, severe muscle weakness including respiratory muscle weakness and dysphagia, and skin manifestations such as necrosis, periungual erythema, and the V or shawl sign are associated with occult malignancy in patients with myositis. Treatment-resistant skin changes (eg, ulcerations and necrosis) are observed more commonly in CAM (44% vs 12%, P <.05). Necrotizing cutaneous lesions and severe pruritus are predictive factors for CAM, and leukocytoclastic vasculitis has been seen in patients with DM and malignancy.
The muscle weakness of CAM is generally rapidly progressive and severe, leading to immobility (39% vs 12%, P <.05) and may be distal in location as well (57% vs 6%, P <.05). Dysphagia (50% vs 36%), oropharyngeal dysfunction (32% vs 13%), and respiratory muscle involvement (32% vs 16%, P <.05) are more frequent in CAM, and respiratory insufficiency due to respiratory muscle involvement can be fatal. Joint symptoms (16% vs 51%, P <.05), Raynaud syndrome (11% vs 26%, P <.05), and fever (0% vs 29%, P <.05) were less common in CAM and may be protective. In a univariate analysis of 121 patients with DM compared with 29 patients with CAM seen over 13 years, male sex, an older age at presentation, the acute onset of cutaneous or muscular symptoms, periungual erythema, necrotic skin lesions, higher creatine kinase (CK) levels, antinuclear antibody titers, and lower C4 and C3 levels were associated with underlying malignancy in DM patients. In the subsequent multivariate analysis, age over 52, rapid onset of skin or muscle symptoms, cutaneous necrosis, periungual erythema, and low C4 levels were the only factors associated with higher risk of CAM.
The presence of interstitial lung disease (ILD) and antisynthetase autoantibodies seems to be protective for the development of cancer, and some report no cases of ILD among patients with cancer in any of the IIM subsets (estimated OR 0.07, 95% CI 0.004–1.18). Earlier studies report similar findings, and many investigators note a lower frequency or absence of anti-Jo-1 autoantibody in the sera of CAM patients. However, case reports of malignancy in patients with ILD and anti-Jo-1 antibodies have surfaced, so a positive Jo-1 autoantibody and ILD do not preclude a consideration of malignancy in certain clinical settings. Moreover, several cases of ILD in CAM patients noted a fulminant course leading to death from respiratory insufficiency before the malignancy work-up was completed.
Although the data indicate that the absence of the more commonly available serologies confers a greater risk of malignancy, autoantibodies are not entirely protective against CAM. A positive antinuclear antibody (ANA) and extractable nuclear antigen were significantly more frequent in idiopathic myositis as compared with CAM. Similarly, positive myositis-specific and myositis-associated antibodies (for example anti-PM-Scl, anti-Ro, antisynthetase antibodies) are less frequently associated with CAM.
Despite the severity of myopathic symptoms in CAM, the serum CK and lactate dehydrogenase (LDH) levels were often less elevated, and may be within the normal range. However, the reverse has also been reported. An interesting association between severe capillary derangement on capillaroscopy and CAM has been found (3 of 6 in CAM vs 3 of 47 in non-CAM patients) (odds ratio [OR] 14.7; 95% CI, 2.0–106.6).
There are no characteristic histopathologic findings in patients with myositis that consistently raise the suspicion of an associated occult malignancy. Nevertheless, necrotizing myopathy with pipestem microvascular changes has been associated with cancer. Muscle biopsy specimens characteristically show evidence of massive necrosis of the muscle fibers with almost complete absence of inflammatory infiltrates, perhaps suggesting involvement of humoral immunologic mechanisms directed against tumor cells.
Risk factors for CAM
Identifying IIM patients at risk of developing malignancy has clear importance to the managing clinician. Older age, male sex, refractory or recurrent myositis, severe muscle weakness including respiratory muscle weakness and dysphagia, and skin manifestations such as necrosis, periungual erythema, and the V or shawl sign are associated with occult malignancy in patients with myositis. Treatment-resistant skin changes (eg, ulcerations and necrosis) are observed more commonly in CAM (44% vs 12%, P <.05). Necrotizing cutaneous lesions and severe pruritus are predictive factors for CAM, and leukocytoclastic vasculitis has been seen in patients with DM and malignancy.
The muscle weakness of CAM is generally rapidly progressive and severe, leading to immobility (39% vs 12%, P <.05) and may be distal in location as well (57% vs 6%, P <.05). Dysphagia (50% vs 36%), oropharyngeal dysfunction (32% vs 13%), and respiratory muscle involvement (32% vs 16%, P <.05) are more frequent in CAM, and respiratory insufficiency due to respiratory muscle involvement can be fatal. Joint symptoms (16% vs 51%, P <.05), Raynaud syndrome (11% vs 26%, P <.05), and fever (0% vs 29%, P <.05) were less common in CAM and may be protective. In a univariate analysis of 121 patients with DM compared with 29 patients with CAM seen over 13 years, male sex, an older age at presentation, the acute onset of cutaneous or muscular symptoms, periungual erythema, necrotic skin lesions, higher creatine kinase (CK) levels, antinuclear antibody titers, and lower C4 and C3 levels were associated with underlying malignancy in DM patients. In the subsequent multivariate analysis, age over 52, rapid onset of skin or muscle symptoms, cutaneous necrosis, periungual erythema, and low C4 levels were the only factors associated with higher risk of CAM.
The presence of interstitial lung disease (ILD) and antisynthetase autoantibodies seems to be protective for the development of cancer, and some report no cases of ILD among patients with cancer in any of the IIM subsets (estimated OR 0.07, 95% CI 0.004–1.18). Earlier studies report similar findings, and many investigators note a lower frequency or absence of anti-Jo-1 autoantibody in the sera of CAM patients. However, case reports of malignancy in patients with ILD and anti-Jo-1 antibodies have surfaced, so a positive Jo-1 autoantibody and ILD do not preclude a consideration of malignancy in certain clinical settings. Moreover, several cases of ILD in CAM patients noted a fulminant course leading to death from respiratory insufficiency before the malignancy work-up was completed.
Although the data indicate that the absence of the more commonly available serologies confers a greater risk of malignancy, autoantibodies are not entirely protective against CAM. A positive antinuclear antibody (ANA) and extractable nuclear antigen were significantly more frequent in idiopathic myositis as compared with CAM. Similarly, positive myositis-specific and myositis-associated antibodies (for example anti-PM-Scl, anti-Ro, antisynthetase antibodies) are less frequently associated with CAM.
Despite the severity of myopathic symptoms in CAM, the serum CK and lactate dehydrogenase (LDH) levels were often less elevated, and may be within the normal range. However, the reverse has also been reported. An interesting association between severe capillary derangement on capillaroscopy and CAM has been found (3 of 6 in CAM vs 3 of 47 in non-CAM patients) (odds ratio [OR] 14.7; 95% CI, 2.0–106.6).
There are no characteristic histopathologic findings in patients with myositis that consistently raise the suspicion of an associated occult malignancy. Nevertheless, necrotizing myopathy with pipestem microvascular changes has been associated with cancer. Muscle biopsy specimens characteristically show evidence of massive necrosis of the muscle fibers with almost complete absence of inflammatory infiltrates, perhaps suggesting involvement of humoral immunologic mechanisms directed against tumor cells.
Autoantibodies associated with CAM
No specific or sensitive clinical or biologic marker for the diagnosis of CAM has been identified. As noted previously, the absence of autoantibodies may predict a higher risk of occult malignancy, whereas the presence of antisynthetase autoantibodies seems to be protective against cancer. To achieve a higher predictive value for diagnosing CAM, Chinoy and colleagues cross-sectionally determined the antibody profile of Caucasian myositis patients from the United Kingdom, specifically assessing anti-Jo-1, anti-PM-Scl, anti-U1-RNP, anti-U3-RNP, and anti-Ku antibodies. They found a paucity of these autoantibodies in CAM patients quantifying a 6- to 7-fold increased relative cancer risk in individuals lacking this panel of autoantibodies compared with those possessing such antibodies. Finding this panel of autoantibodies negative had a high negative predictive value for CAM, perhaps helpful to managing clinicians until better predictive markers are commercially available. However, other authors have not confirmed these findings, perhaps related to examining fewer cases, which limited the statistical power.
Anti-Mi-2 is associated with skin changes and possible occult malignancy in some patients with DM. A large study of white patients with myositis found a low prevalence of anti-Mi-2 antibodies in CAM, while a European study suggested that cancer risk was increased, but only in anti-Mi-2 antibody-positive patients possessing the N-terminal fragment of the Mi-2 antigen.
A newly described autoantibody (anti-p155) against a 155-kDa nuclear protein, identified as transcription intermediary factor 1-gamma (TIF1-g), has proven useful for cancer screening in patients with DM. This marker is also found in JDM but is not associated with cancer in this IIM subset as opposed to adults having an increased risk of CAM who possess this autoantibody. A recent study proposed that anti-p155 is involved in the transforming growth factor beta (TGF-β) signaling pathway, which is inactivated in some malignancies, thus linking TIF1-γ to carcinogenesis. The association between adult CAM and anti-p155 has been confirmed in other case series. Possession of the anti-p155 autoantibody represents a significant risk factor for CAM, and it was found exclusively in DM with a high negative predictive value for CAM. Its role in PM is unclear and requires further investigation. The value of anti-p155 to predict the existence of occult cancer seems to be limited to DM patients, the most highly affected myositis subset for cancer.
In a recent meta-analysis of six published reports, anti-p155 was associated with cancer, with an overall specificity of 70%, sensitivity of 89%, negative predictive value (NPV) of 93%, and a diagnostic odds ratio of 18 (95% CI 8–40). Thus, 70% of myositis patients with anti-p155 will develop cancer, and positive testing leads to an 18-fold higher association with cancer than anti-p155 negative status. Thus, this autoantibody is potentially very helpful in the work-up and follow-up of myositis patients. Moreover, additional studies suggest that ILD and positive antisynthetase autoantibodies or the presence of anti-Mi-2 (subunit Mi-2b), previously reported to be protective or associated with cancer, respectively, were congruent with positive or negative anti-p155 autoantibodies.
Chinoy and colleagues combined the results of the aforementioned autoantibody panel (anti-Jo-1, anti-Ku, anti-PM-Scl, anti-U1-RNP, anti-U3 – RNP) and anti-p155 testing in a cohort of 282 myositis patients to predict CAM. They showed that a negative autoantibody panel alone was highly sensitive (87%) and had high NPV (98%), but low specificity (45%) for CAM. A positive anti-p155 alone was 50% sensitive for the detection of CAM with a 42% positive predictive value (PPV) (ie, 58% false-positive rate). However, the anti-p155 antibody test was 96% specific (most non-CAM patients were negative for this antibody) and demonstrated a high NPV (97% of patients without anti-p155 antibody did not have CAM). Finally, when the autoantibody panel and anti-p155 were negative, there was 94% sensitivity and 99% NPV (indicating that only 1% with a negative autoantibody panel or a negative anti-p155 had CAM). When the DM group was analyzed alone, this combined strategy yielded 100% sensitivity and NPV. However, the specificity and PPV of the combined strategy was low, at 45% and 9%, respectively, indicating these tests are most helpful to rule out malignancies, since positive anti-p155 autoantibody results can be seen in many patients without CAM.
The factors that lead to the development of cancer in some patients with this autoantibody remain unclear and warrants further investigation. It is possible that different epitopes or varying titers of anti-p155 lead to different clinical risk and phenotypes. Moreover, longitudinal measurement of these autoantibodies may provide further information on risk and prognosis of cancer and myositis in patients possessing this marker. Commercial testing of this autoantibody is not yet available.
Related posts:
Paraneoplastic and Cancer Treatment-Related Rheumatic Disorders
Neoplastic and Paraneoplastic Synovitis
Rheumatic Manifestations of Primary and Metastatic Bone Tumors and Paraneoplastic Bone Disease
Neoplastic and Paraneoplastic Vasculitis, Vasculopathy, and Hypercoagulability
Neoplasm Mimics of Rheumatologic Presentations: Sialadenitis, Ocular Masquerade Syndromes, Retroperitoneal Fibrosis, and Regional Pain Syndromes
Neoplastic/Paraneoplastic Dermatitis, Fasciitis, and Panniculitis
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
