Abstract
This cohort study estimates the prevalence of nonretroperitoneal abdominal organ involvement in Erdheim-Chester disease in a large cohort of patients.
Erdheim-Chester disease (ECD) is a multisystem tissue accumulation of foamy histiocytes with constant activation of the mitogen-activated protein kinase pathway leading to organ dysfunction. Specific imaging features, such as perirenal fat stranding, bilateral osteosclerosis of the long bones, parietal aortic-arch sheathing, and sclerosing facial sinuses are well described, though very few descriptions of the involvement of abdominal organs outside the retroperitoneum are reported.1,2,3 The aim of this study was to estimate the prevalence of nonretroperitoneal abdominal organ involvement in ECD in a large cohort of patients.
Methods
Written informed consent was obtained for all patients when included in the French Histiocytosis registry for retrospective data analysis on their imaging and electronic health records with the approval of the Comite de Protection des Personnes Sud-Ouest Et Outre-Mer II ethics committee (ID RCB:2019-A01814-53). The STROBE reporting guideline was followed.
In this cohort study, we retrospectively studied 304 patients who were 18 years or older and referred to the French national reference center for histiocytosis between September 2002 and March 2021. Among them, 252 (83%) had a histologically proven diagnosis of pure ECD, and 52 (17%) had a variant ECD (ECD plus another histiocytosis).3,4 The computed tomography (CT) scans performed closest to the date of histopathological diagnosis were analyzed retrospectively by 2 abdominal radiologists (B.D. and O.L.). Imaging pattern frequencies were assessed and compared between pure and variant ECD and between patients with and without BRAF V600E variations with χ2 tests with or without Yates correction. Kaplan-Meier analysis with log-rank tests analyzed the association of relevant imaging patterns with overall survival. All tests were 2-tailed, with P values less than .05 indicating statistical significance.
Results
In this cohort of 304 patients (mean [range] age, 60 [23-90] years; 211 men [69.4%]), the male/female patient ratio was 3:1 for pure ECD and 1:1 for variant ECD (χ2: P = .003). Nonretroperitoneal imaging findings are summarized in the Table. All numbers are single observed proportions with symmetrical plus/minus 95% CIs to estimate the theoretical proportions for each imaging pattern. The most common finding on CT scans was mesenteric panniculitis (MP), observed in 58 of 304 patients (19.1% ± 4.4%) (Figure). This pattern was associated in 52 of 58 (89.7% ± 3.3%) cases with perirenal fat stranding, a quasi-pathognomonic retroperitoneal finding of ECD. A mesenteric fat biopsy with histological findings of ECD was available for 13 patients, 9 of whom presented with typical MP on CT scan. Other imaging findings among the 304 patients were splenomegaly in 34 (11.2% ± 3.5%), hepatomegaly in 25 (8.2% ± 3.1%), noncirrhotic dysmorphia in 19 (6.2% ± 2.7%), perfusion disorders in 17 (5.6% ± 2.6%), and sclerosing cholangitis pattern in 9 (3.0% ± 1.9%). The prevalence of CT signs did not differ between the pure and variant ECD groups. The condition MP was statistically significantly more present in patients with BRAF V600E variation (38 of 145, 26.2% ± 7.1%) than in patients with BRAF V600E wild type (7 of 86, 8.1% ± 5.8%; χ2: P = .001). The overall mortality rate was 90 of 304 patients (29.6% ± 5.1%). There was no statistically significant difference in survival between pure and variant ECD groups. Survival was only statistically significantly shorter in patients with hepatomegaly (median [SD] overall survival, 62.7 [17.1] vs 93.9 [10.2] months; P = .03; odds ratio, 1429 [95% CI, 1023-1995]).
Table. Main Imaging Findings on Computed Tomography (CT) Scans.
| Organ/system | Imaging findings | No. (% ± 95% CI)a | P valueb | ||
|---|---|---|---|---|---|
| All patients | ECD | ||||
| Pure | Variant | ||||
| No. | 304 | 252 | 52 | NA | |
| Mesentery | Mesenteric panniculitis | 58 (19.1 ± 4.4) | 51(20.2 ± 5.0) | 7 (13.4 ± 9.3) | .33 |
| Liver | Dysmorphia (cirrhosislike) | 1 (0.3 ± 0.6) | NA | 1 (1.9 ± 3.7) | NA |
| Dysmorphia (noncirrhotic) | 19 (6.2 ± 2.7) | 16 (6.3 ± 3.0) | 3 (5.8 ± 6.3) | .80c | |
| Hepatomegaly | 25 (8.2 ± 3.0) | 21 (8.3 ± 3.4) | 4 (7.7 ± 7.2) | .90c | |
| Rounded edges | 6 (2.0 ± 1.6) | 4 (1.6 ± 1.5) | 2 (3.8 ± 5.2) | .75c | |
| Steatosis | 13 (4.3 ± 2.3) | 12 (4.8 ± 2.6) | 1 (1.9 ± 3.7) | .70c | |
| Perfusional disorders | 17 (5.6 ± 2.6) | 13 (5.2 ± 2.7) | 4 (7.7 ± 7.2) | .90c | |
| Portal hypertension | 9 (3.0 ± 1.9) | 8 (3.1 ± 2.2) | 1 (1.9 ± 3.7) | .99c | |
| Pedicle adenomegaly | 11 (3.6 ± 3.0) | 8 (3.1 ± 2.2) | 3 (5.8 ± 6.3) | .80c | |
| Hypovascular nodules | 2 (0.6 ± 1.7) | 1 (0.4 ± 0.7) | 1 (1.9 ± 3.7) | NA | |
| Hypervascular nodules | 6 (2.0 ± 1.6) | 6 (2.4 ± 1.9) | NA | NA | |
| Veins | SV thrombosis | 2 (0.6 ± 1.7) | 2 (0.8 ± 1.1) | NA | NA |
| Biliary tree | Sclerosing cholangitis pattern | 9 (3.0 ± 1.9) | 7 (2.8 ± 2.0) | 2 (3.8 ± 5.2) | .99c |
| Gall bladder wall thickening | 6 (2.0 ± 1.6) | 5 (2.0 ± 1.7) | 1 (1.9 ± 3.7) | .99c | |
| Pancreas | Diffuse dilation of MPD | 4 (1.3 ± 1.3) | 4 (1.6 ± 1.5) | NA | NA |
| Stenosis of MPD | 1 (0.3 ± 0.6) | 1 (0.4 ± 0.7) | NA | NA | |
| Dilation of APD | 10 (3.2 ± 2.0) | 9 (3.6 ± 2.3) | 1 (1.9 ± 3.7) | .80c | |
| Pancreatic mass | 1 (0.3 ± 0.6) | 1 (0.4 ± 0.7) | NA | NA | |
| Spleen | Splenomegaly (>13 cm) | 34 (11.2 ± 3.5) | 28 (11.1 ± 3.9) | 6 (11.5 ± 8.7) | .95 |
| Perisplenic fat stranding | 3 (0.9 ± 2.1) | 2 (0.8 ± 1.1) | 1 (1.9 ± 3.7) | .99c | |
| Spleen nodules | 9 (3.0 ± 1.9) | 7 (2.8 ± 2.0) | 2 (3.8 ± 5.2) | .99c | |
| Bowels | Wall thickening of the bowel | 1 (0.3 ± 0.6) | NA | 1 (1.9 ± 3.7) | NA |
Abbreviations: APD, accessory pancreatic ducts; ECD, Erdheim-Chester disease; MPD, main pancreatic duct; NA, not applicable; SV, splenic vein.
All numbers are single observed proportions with symmetrical plus/minus 95% CIs to estimate the theoretical proportions for each imaging pattern.
P values were calculated with a χ2 test with Yates correction in case of small data.
Yates correction was used in these cases because of small data sets.
Figure. Computed Tomography (CT) Patterns of Biopsy-Proven Erdheim-Chester Disease (ECD)–Related Mesenteric Panniculitis.
Mesenteric panniculitis seen (white arrowheads) in a patient in their 70s (A) and a patient in their 60s (B). Both had pure ECD. Perirenal fat stranding features (yellow arrowheads) were also visible.
Discussion
In this cohort study, which, to our knowledge, represented the largest ECD cohort studied to date, we found that MP was the most common finding of nonretroperitoneal involvements seen in 19.1% ± 4.4% of patients, though generally observed in 2% of patients’ CT scans in the general population.4 Only a few descriptions of the association of MP with ECD have been reported so far.5,6 Splenomegaly was the second most common condition encountered in 11.2% ± 3.5%. Sclerosing cholangitis was described in only 3.0% ± 1.9%. In the present study, we found sclerosing cholangitis patterns in 9 of 304 (3.0%) patients. Hepatomegaly was the only condition associated with shorter survival.
This study has some limitations. It was a retrospective study resulting in technical heterogeneity of CT scans, although this was unavoidable because ECD is too rare for a prospective study. Moreover, the analysis of CT scans alone was not optimal for identifying hepatic nodules or biliary duct disease, which may have been underestimated. Nevertheless, these limitations do not impact the results obtained for MP.
In this cohort study of 304 patients, focusing on nonretroperitoneal abdominal involvement in biopsy-proven ECD, we identified a high prevalence, until now never reported to our knowledge, of mesenteric panniculitis, also significantly associated with the BRAF V600E variation.
References
- 1.Wang F, Cao X, Niu N, et al. Multisystemic imaging findings in Chinese patients with Erdheim-Chester disease. AJR Am J Roentgenol. 2019;213(6):1179-1186. doi: 10.2214/AJR.19.21523 [DOI] [PubMed] [Google Scholar]
- 2.Young JR, Johnson GB, Murphy RC, Go RS, Broski SM. 18F-FDG PET/CT in Erdheim-Chester disease: imaging findings and potential BRAF mutation biomarker. J Nucl Med. 2018;59(5):774-779. doi: 10.2967/jnumed.117.200741 [DOI] [PubMed] [Google Scholar]
- 3.Pan A, Doyle T, Schlup M, Lubcke R, Schultz M. Unusual manifestation of Erdheim-Chester disease. BMC Gastroenterol. 2011;11(1):77. doi: 10.1186/1471-230X-11-77 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Badet N, Sailley N, Briquez C, Paquette B, Vuitton L, Delabrousse É. Mesenteric panniculitis: still an ambiguous condition. Diagn Interv Imaging. 2015;96(3):251-257. doi: 10.1016/j.diii.2014.12.002 [DOI] [PubMed] [Google Scholar]
- 5.Perlat A, Decaux O, Sébillot M, Grosbois B, Desfourneaux V, Meadeb J. Erdheim-Chester disease with predominant mesenteric localization: lack of efficacy of interferon alpha. Joint Bone Spine. 2009;76(3):315-317. doi: 10.1016/j.jbspin.2008.09.013 [DOI] [PubMed] [Google Scholar]
- 6.Lamboley JL, Le Moigne F, Felten D, Crozes C, Farthouat P, Pavic M. Atteinte mésentérique d’une maladie d’Erdheim-Chester : une localisation rare: Erdheim-Chester disease with mesenteric involvement. J Radiol. 2010;91(7-8):805-808. doi: 10.1016/S0221-0363(10)70119-0 [DOI] [PubMed] [Google Scholar]