The New Histopathologic Classification of ANCA-Associated GN and Its Association with Renal Outcomes in Childhood

Abstract

Background and objectives

A proposed histopathologic classification for ANCA-associated GN is predictive of long-term renal outcome in adult populations. This study sought to validate this system in a pediatric cohort.

Design, setting, participants, & measurements

This was a retrospective, single-center, cohort study of 40 children diagnosed and followed until their transition to adult care at one institution between 1987 and 2012. Renal biopsy specimens were reviewed by a pathologist blinded to patient outcome and were classified using the new histopathologic classification system of focal, crescentic, mixed, and sclerotic groups. Time to the composite outcome of CKD stages 3 and 4 (determined by eGFR with repeated creatinine measures using the Schwartz equation) or ESRD (defined as dialysis dependence or transplantation) were ascertained.

Results

The study population consisted of 40 children (70% female), followed for a median of 2.4 years. The biopsy specimens were categorized as focal in 13 patients (32.5%), crescentic in 20 (50%), mixed in two (5%), and sclerotic in five (12.5%). Mixed and crescentic were combined for analyses. Survival analysis of time to the composite renal endpoint of at least 3 months of eGFR<60 ml/min per 1.73 m² or ESRD differed significantly among the three biopsy groups log-rank P<0.001), with an adjusted hazard ratio of 3.14 (95% confidence interval, 0.68 to 14.4) in the crescentic/mixed group and 23.6 (95% confidence interval, 3.9 to 144.2) in the sclerotic category compared with the focal category. The probability of having an eGFR>60 ml/min per 1.73 m² at 2 years was 100% for the focal, 56.5% for the crescentic/mixed, and 0% for the sclerotic biopsy categories.

Conclusions

This study showed the clinical utility of this histopathologic classification system and its ability to discriminate renal outcomes among children with ANCA GN.

Introduction

ANCA-associated vasculitides (AAV) include granulomatosis with polyangiitis (GPA), microscopic polyangiitis (MPA), and eosinophilic granulomatosis with polyangiitis. Although AAV are rare in childhood, renal involvement is common, occurring in 60%–90% of reported pediatric cases (1,2), and are most frequently associated with MPA (3). The most severe form of renal involvement in AAV is the rapidly progressive ANCA-associated necrotizing crescentic GN; among children, 30%–40% will progress to CKD and about 10% to ESRD by adulthood (2,4).

In 2010, an international working group of renal pathologists developed a new histopathologic classification system for ANCA GN and validated it in a population of 100 adults with ANCA GN from two European Vasculitis Study Group trials (5–7). Preceding that, various pathologic criteria, including percentage of normal glomeruli, degree of chronic (sclerotic) lesions, and active (crescentic) lesions predicted kidney outcomes in several studies (8–10). Collating these data, the working group proposed a novel and intuitive schema classifying the renal histopathologic features into four distinct categories: focal, crescentic, mixed, and sclerotic. The biopsy category is reliable in predicting long-term renal outcome (eGFR) at both 1 and 5 years (5).

Any newly devised classification system needs external validation in the pediatric population to confirm both relevance and reproducibility as a prognostic tool, especially because pediatric populations may differ in both disease severity and outcomes. For example, when the predictive value of the Oxford classification system for IgA nephropathy was applied to children, there was inconsistent association of the pathologic criteria with renal outcomes (11–15). Therefore, we sought to validate this new pathologic classification of ANCA GN in cases with childhood onset and determine the association with renal outcomes.

Materials and Methods

This was a nonconcurrent, single-center cohort study of consecutive children diagnosed with pauci-immune necrotizing GN in association with granulomatosis with GPA or MPA at the Hospital for Sick Children. A complete population was ascertained as previously described; the study included 24 new patients and extended follow-up of the 16 patients described in an earlier study (2). Children were diagnosed and followed up until their transfer to an adult center from January 1, 1987, to August 31, 2012. All patients met classification criteria of the American College of Rheumatology or Chapel Hill Classification systems for GPA or MPA and satisfied the European League Against Rheumatism and the Pediatric Rheumatology European Society criteria for the classification of childhood vasculitides (16). The study included children age≤18 years of age who had an initial renal biopsy at the Hospital for Sick Children. Patients were excluded if they had an overlap syndrome, such as anti-glomerular basement membrane/ANCA-positive vasculitis, or the biopsy specimen was unavailable or deemed inadequate for histopathologic classification by the pathologist. The Research Ethics Board at the Hospital for Sick Children approved the study, and we adhered to the Declaration of Helsinki.

Data Collection

All data were collected from the electronic patient record or the archived clinical chart. Data were collected on demographic characteristics, clinical symptoms, laboratory data, and kidney function. eGFR was calculated using the modified Schwartz equation (17). Classification of CKD grade did not differ with the change of creatinine assay to an isotope dilution mass spectrometry reference method in 2008. Treatment data included use of plasma exchange, antihypertensive medications or angiotensin-converting enzyme inhibitors, and type of immunosuppressive therapy used (including steroids, cyclophosphamide, methotrexate, mycophenolate mofetil, and rituximab).

Renal Histopathology

All renal biopsy specimens were re-reviewed by a single pediatric renal pathologist, blinded to patient outcome. For light microscopy, all biopsy specimens had been fixed in 10% formalin, routinely processed, and sectioned at 2 microns. Fifteen levels were stained on each specimen; three each were stained with hematoxylin and eosin, hematoxylin-phloxine-safranin, periodic acid–Schiff, periodic acid ammoniacal silver, and Masson trichrome. Biopsy specimens were then classified using the published histopathologic schema (5) as focal, crescentic, mixed, or sclerotic on the basis of the light microscopic findings (Figure 1).

Figure 1.

Illustration of the histopathological scoring system for ANCA-associated GN. Glomeruli were categorized as normal (A and B), those with cellular crescents (C and D), or sclerotic (E and F). The left column (A, C, E) illustrates a representative glomerulus in each category, stained with hematoxylin and eosin. In the right column (B, D, F), glomeruli have been stained with periodic acid–Schiff to highlight the basement membranes in the vascular tuft, making the crescent easier to distinguish from the glomerulus in the case of the cellular crescent, and to demonstrate the obliteration of the glomerulus in the sclerotic category. (A–F, original magnification, ×400.)

Outcomes

The primary outcomes measured were (1) time to the composite outcome of CKD stages 3 and 4 (defined as an eGFR 15–59 ml/min per 1.73m²) or ESRD (defined as having an eGFR of <15 ml/min per 1.73 m²), dialysis dependence or transplantation; (2) time to CKD stages 3 and 4; and (3) the percentage of patients reaching CKD or ESRD at 6 months and longest follow-up. Secondary outcomes included degree of proteinuria, presence of hypertension, and number of relapses (defined as an increase in immunosuppressive treatment due to renal disease, suggested by an increase in serum creatinine or proteinuria and confirmed by renal biopsy). Data were obtained at 6 months and last follow-up. In addition, all serum creatinine measures from diagnosis to last follow-up were collected from the electronic patient record.

Statistical Analyses

Summary statistics were used to summarize patient demographic characteristics and clinical features at presentation. Normally distributed data are presented as mean±SD; otherwise, medians with interquartile range are provided. We presented the mean eGFR per 1-month period over an 18-month follow-up time. The mean number of distinct creatinine values per patient was 94 (range, 7–753). Urinary protein-to-creatinine ratios and eGFR for the remainder of analyses were log transformed. Comparisons are made of baseline information by biopsy category using appropriate tests for continuous or categorical measures.

We determined time to CKD by eGFR with repeated creatinine measures using the Schwartz equation, or ESRD as defined as dialysis dependence or transplantation. Kaplan–Meier survival of time to ESRD or the composite outcome of CKD (eGFR< 60 ml/min per 1.73 m²) or ESRD and linear regression analyses of the slope of eGFR were performed. All patients who received dialysis therapy at dialysis initiation were set to ESRD if they remained on maintenance dialysis for at least 3 months. We also conducted a Cox regression analysis to determine the risk of biopsy category and CKD/ESRD. A P value of 0.05 indicated a statistically significant difference. Data were analyzed using Stata software, version 11.

Results

The study population consisted of 40 children with ANCA GN seen from January 1, 1987, to August 31, 2012, consecutively. There were an equal number of patients with GPA and MPA. The median age of the patients at renal diagnosis was 12 (range, 3.6–17.3) years, and 70% of the cohort was female.

Renal Pathology

Upon classification as per the new schema (outlined in Table 1), there were 13 (32.5%) cases in the focal category, 20 (50%) in the crescentic category, two (5%) in the mixed category, and five (12.5%) in the sclerotic category. For the remainder of the analyses, we combined the two children with mixed pathologic features into the crescentic group (Table 2). The eGFR at baseline in the mixed group ranged from 87–138.7 ml/min per 1.73 m². By definition, patients in the mixed group had <50% normal, <50% crescentic, and <50% globally sclerotic glomeruli. A glomerulus was described as being sclerotic when >80% of the glomerulus was sclerosed. Across categories from focal to sclerotic, serum creatinine, albumin, eGFR, hemoglobin, erythrocyte sedimentation rate, urine protein-to-creatinine ratio and need for dialysis at baseline varied significantly.

Table 1.

Pathologic classification schema

Biopsy Category	Definition
Focal	≥50% normal glomeruli
Crescentic	≥50% glomeruli with cellular crescents
Mixed	<50% normal, <50% crescentic, <50% globally sclerotic glomeruli
Sclerotic	≥50% globally sclerotic glomeruli

Data obtained from reference 5.

Table 2.

Baseline demographic and clinical characteristics of 40 children with ANCA GN from 1987 to 2012

Characteristic	Focal ( n=13)	Crescentic/Mixed (n=22)	Sclerotic (n=5)	P Value
Demographic
Male	2	9	1	0.74
Age at ANCA GN diagnosis (yr)	13.5 (7–15)	11 (8–15)	11.2 (11.1–13.2)	0.84
Time to ANCA GN Diagnosis (d)	109 (34–163)	26 (10–51)	24 (14–86)	0.16
Diagnosis
MPA	4 (30.7)	13 (59.1)	4 (80)	0.05
GPA	8 (66.7)	10 (43.5)	1 (20)	0.05
ANCA positive	9 (81.8)	18 (85.7)	4 (80)	0.79
ANCA specificity
p-ANCA	4 (36.4)	11 (57.9)	4 (80)	0.29
c-ANCA	5 (45.5)	6 (31.6)	0	—
PR3	5 (55.5)	10 (62.5)	3 (100)	0.37
MPO	4 (44.5)	6 (37.5)	0 (0)
ELISA titer
1/80	2 (22.2)	3 (17.7)	0	0.76
1/160	2 (22.2)	6 (35.3)	1 (25)	–
1/320	3 (33.3)	2 (11.8)	1 (25)	–
1/640	2 (22.2)	6 (35.3)	2 (50)	–
Kidney status at presentation
Creatinine (mg/dl)	0.6 (0.5–0.9)	2.8 (1.1–6.4)	3.5 (3.1–9.7)	0.01
eGFR (ml/min per 1.73 m2)	106.3 (84.8–126.6)	19.6 (10.7–57.7)	19.4 (15–20.0)	<0.001
Proteinuria non-nephrotic (n)	10	12	1	0.01
Nephrotic (n)	1	8	2	–
Protein-to-creatinine ratio (mg/g)	695 (265–1425)	1894 (1327–3735)	24,925 (13,000–36,849)	0.02
Need for dialysis at diagnosis	0	9 (39)	3 (60)	0.01
Other
Hemoglobin (g/dL)	11.7 (11.0–11.8)	9.85 (8.1–10.8)	8.5 (6.5–9.3)	0.02
Platelets (×109/L)	291 (240–355)	361 (294–416)	391 (238–438)	0.42
Neutrophils (×109/L)	7.4 (4.4–8.8)	9.7 (6.8–12.6)	5.5 (4.4–9.6)	0.23
Albumin (g/dl)	3.6 (3.5–3.9)	3.1 (2.7–3.4)	2.6 (2.6–3.1)	0.01
ESR (mm/hr)	32 (12–90)	103 (54–133)	93 (78–110)	0.02
CRP (mg/L)	3.4 (0.9–22.2)	55.5 (7.3–183)	59.3 (0.6–118)	0.32
C3 (g/L)	1.2 (0.99–1.3)	1.3 (1.1–1.5)	0.97 (0.85–1.0)	0.39
IgA (mg/dl)	150 (120–200)	150 (100–280)	220 (130–290)	0.71
IgM (mg/dl)	120 (80–170)	90 (60–130)	125 (120–150)	0.53
IgG (mg/dl)	1050 (720–1330)	990 (620–1490)	1300 (920–1590)	0.79
Treatment of ANCA GN
Corticosteroids	11 (84.6)	22 (100)	5 (100)	0.14
Cyclophosphamide	5 (38.5)	17 (77.3)	2 (40)	0.05
Plasma exchange	1 (7.7)	4 (18.1)	1 (10)	0.68
Azathioprine	2 (15.4)	0	0	0.12
Methotrexate	4 (30.8)	0	0	0.03
Rituximab	2 (15.4)	0	0	0.12

Continuous variables were analyzed using one-way ANOVA and categorical variables by chi- square analyses. Unless otherwise noted, values are expressed as number (percentage) of patients or median (range). MPA. microscopic polyangiitis, GPA, granulomatosis with polyangiitis; p-ANCA, perinuclear ANCA; c-ANCA, cytoplasmic ANCA; PR3, proteinase 3; MPO, myeloperoxidase; ESR, erythrocyte sedimentation rate; CRP, C-reactive protein.

Outcomes

The median duration of follow-up was 883 (interquartile range, 95–1480) days. Four of the 13 (31%) patients requiring dialysis at diagnosis were dialysis independent at 6 months. Mean duration of dialysis in those that came off dialysis was 9.7 days (range, 7–30). No patients died during the study period.

We conducted a Kaplan–Meier survival analysis of time to the composite renal endpoint of at least 3 months of eGFR<60 ml/min per 1.73 m² or ESRD (dialysis or transplantation) by the three biopsy groups (log-rank P <0.001). We limited follow-up to 2 years as seven children transitioned to adult care. Probability (95% confidence interval [95% CI]) of eGFR>60 ml/min per 1.73 m² at 2 years after diagnosis was 100% for focal, 56.5% (95% CI, 32.6% to 74.7%) for crescentic/mixed, and 0% for sclerotic biopsy categories (Figure 2A). Probability of not reaching ESRD (dialysis or transplantation) at 2 years after diagnosis was 100% for focal, and 67.3% (95% CI, 43.2% to 82.9%) for crescentic/mixed. Those with sclerosis all progressed to ESRD (Figure 2B).

Figure 2.

Renal survival by biopsy category. (A) Time to composite outcome of CKD/ESRD and probability of renal survival at 2 years after diagnosis of ANCA GN. Probability (95% confidence interval [95% CI]) of renal survival at 2 years after diagnosis by kidney biopsy category. Survival functions for all biopsy categories differed significantly (log-rank P<0.001). (B) Time to ESRD and probability of renal survival at 2 years after diagnosis of ANCA GN. Probability (95% CI) of eGFR>60 ml/min per 1.73 m² (excluding patients who underwent transplantation) at 2 years after diagnosis by kidney biopsy category. Survival functions for all biopsy categories differed significantly (log-rank P<0.01).

By Cox regression analyses, we demonstrated that children in the crescentic/mixed groups were three times more likely to develop the composite endpoint of CKD/ESRD (adjusted hazard ratio, 3.14; 95% CI, 0.68 to 14.4); however, the confidence interval crossed 1. Those with sclerosis were at 23.6 times higher risk than those with focal pathology (95% CI, 3.9 to 144.2) while controlling for age at diagnosis and sex. On the basis of the small sample size, we had wide CIs; however, there was a consistently higher risk for progression in both the crescentic and sclerotic group than in the focal group (P for trend <0.001). Age at diagnosis and sex were not associated with higher risk. Additionally, we could not adjust for urinary protein-to-creatinine ratio or eGFR because they were tightly correlated (R=−0.62; P<0.001). For every increase in the natural log of urinary protein-to-creatinine, there was a higher risk of progression (adjusted hazard ratio of 5.36; 95% CI, 1.99 to 14.42) while controlling for age at diagnosis and log of eGFR at baseline. Similarly, for every increase in the natural log of eGFR at baseline, there was a 90% lower risk of progression to CKD and ESRD (adjusted hazard ratio, 0.12; 95% CI, 0.04 to 0.39) after adjustment for age and urinary protein-to-creatinine ratio. All models were limited by small sample size and thus were minimally adjusted. Any effect related to therapy could not be assessed.

None of the focal group developed CKD stages 3 or 4 or ESRD by 6 months or longest follow-up (Table 3); however, two focal patients had a reduced eGFR of 60–90 ml/min per 1.73 m² at last follow-up but did not meet CKD criteria as defined by stages 3 and 4 CKD. By last follow-up, within the mixed combined category, both children with mixed pathologic type had normal renal function. Among the 20 participants with crescentic pathologic features alone, eight had an eGFR> 60 ml/min per 1.73 m², 3 had CKD stages 3 and 4, and nine had reached ESRD.

Table 3.

Outcomes of 40 children with ANCA GN from 1987 to 2012

Outcome	Focal (n=13)	Crescentic/Mixed (n=22)	Sclerotic (n=5)
6-mo follow-up
eGFR (ml/min per 1.73 m2)	109.9 (100.6–115.4)	57.8 (7.8–97)	10.4 (7.5–11.5)
CKD (eGFR < 60 ml/min per 1.73 m2)	0	4	0
ESRD	0	6	5
Urine protein-to-creatinine ratio (mg/g)	162 (130–212)	1261 (673–2832)	951 (248–2456)
Last follow-up
Duration of follow-up (mo)	71.5 (19.5–91.5)	28.5 (25–47)	79 (42.5–86.5)
eGFR (ml/min per 1.73 m2)	106.2 (99.9–116.1)	72.4 (32.1–94.5)	6.8 (5.7–87.3)
CKD (eGFR < 60 ml/min per 1.73 m2)	0	3	0
ESRD	0	9	5
Urine protein-to-creatinine Ratio (mg/g)	236 (73–381)	235 (142–820)	–

Values are expressed as the number of patients or the median (interquartile range).

Mean eGFR (95% CI) variation over time by biopsy category is presented in Figure 3. By histopathology, there is a clear separation of the mean eGFR for each month over the 18 months since diagnosis. Those in the focal or crescentic/mixed group had an increase in eGFR over the first 3 months after diagnosis, with a slow decline evident in the crescentic/mixed group and no decline evident in the focal group. Linear regression analysis demonstrated an association with slope of (ln transformed) eGFR for the combined crescentic/mixed categories with baseline (ln transformed) protein-to-creatinine ratio (P=0.05) and need for dialysis at presentation (P=0.05) after adjustment.

Figure 3.

Mean eGFR (95% CI) variation over time by biopsy category. With use of repeated measures of serum creatinine to calculate eGFR, the mean eGFR per 1-month period was calculated and plotted over an 18-month follow-up period.

Clinical outcomes at the end of follow-up, and before transition to adult care, demonstrated that 42.4% of the overall cohort had CKD and 33.3% had progressed to ESRD. Among all children, 36.7% were hypertensive and half were prescribed angiotensin-converting enzyme inhibition for proteinuria. Nine (24.3%) children in total were dialysis dependent from diagnosis, and six underwent transplantation (without recurrence) during the study period.

Among a subset of children with focal and crescentic disease, biopsies were repeated to determine progression or because of relapses. Among children with crescentic disease who had a repeat biopsy (n=10), eight progressed to sclerotic disease out of the original 20 children with crescentic disease. Among the focal group, only one had repeat biopsy and was reclassified as mixed.

Median time to diagnosis of AAV from time of symptoms was 15 days (range, 32–109 years). Children with renal AAV almost uniformly received steroids (95%), and 32 children received additional therapies, including rituximab (n=3), methotrexate (n=4), cyclophosphamide (n=26), azathioprine (n=3), or plasma exchange (n=6). Clinical management varied according to the year of diagnosis with availability of steroid-sparing medications, consensus guidelines, and physician practice. Only six children received plasma exchange, with a mean eGFR of 20.3 (range, 6–56) ml/min per 1.73 m²; however, the specific indication for initiation of plasma exchange was difficult to ascertain retrospectively.

Discussion

Childhood-onset ANCA GN poses a significant risk for progressive disease. In our study, 42.4% of children with ANCA GN progressed to CKD, and 33.3% progressed to ESRD in a median of 2.4 years. Our data support the use of this histopathologic classification schema to provide important prognostic information in childhood-onset ANCA GN. By histopathologic classification, patients with sclerotic disease on biopsy fared poorly, with rapid progression to ESRD in less than 6 months. Those children with focal disease presented with mild disease and had a higher baseline eGFR. The children with crescentic/mixed disease had initial improvement in eGFR and a slow decline over 2 years. Half of those with crescentic disease had a renal relapse, and this was associated with progression to sclerotic category on rebiopsy in 80%. Baseline eGFR was an additional predictor for ESRD in the original study of ANCA GN in adults; however, proteinuria was not included in the study (5). We were also able to demonstrate that within the crescentic/mixed category, the need for dialysis at diagnosis of ANCA GN and degree of proteinuria predicted a greater decline in renal function.

The pathologic classification initially developed and validated in adults can now be extended for use in children. The classification not only can predict discrete outcomes of ESRD but also can delineate GFR decline over follow-up and risk of CKD. Thus, at diagnosis, important prognostic information is available for counseling of children and families. Currently, therapies are decided on the basis of extrarenal manifestations and crescentic disease; however, it may be that further discerning protocols based on pathologic features can be considered. A most striking example is that all children in the focal category on biopsy had an eGFR>60 ml/min per 1.73 m², whereas all sclerotic patients progressed to ESRD despite aggressive early therapy. We also observed that patients initially in the crescentic category who relapsed invariably progressed to the sclerotic category.

Substantial morbidity is associated with the treatments for AAV occurring in childhood, including adverse effects of steroids and an increased rate of infections (7-fold increase with steroid and cyclophosphamide used in combination), as well as the more long-term risks of infertility (1,4). Our data suggest that for the sclerotic category, intensive immunosuppressive therapy is unlikely to result in renal function recovery.

Several validation studies of the new pathologic criteria have been published and have demonstrated reliability in predicting outcomes for patients with focal or sclerotic disease in ANCA GN (18–21). One consistent finding, however, in contrast to the original study, has been that the mixed category confers better renal survival than does crescentic disease (5,19,20). We have a small number of children with mixed pathologic features in contrast to all the adult studies but a similar number with crescentic pathologic features (50%); this is more in keeping with the European and Chinese cohorts (47%–55%) and dissimilar to findings in Japan, where only 8% of the population had crescentic disease. This could reflect varying disease processes as all the patients in the Japanese study had antibodies to myeloperoxidase (5,19).

Although outcomes for patients with ANCA GN have improved substantially over the last decade, attributable to earlier diagnosis and better therapeutic strategies, renal involvement in AAV is common, occurring in about 88% of childhood cases and 80% of adult cases (2,20,22). Compared with reported studies of children with ANCA GN, we demonstrated a higher proportion of children reaching CKD or ESRD. We have extended a prior study conducted at the Hospital for Sick Children that found 40% of the first 22 children with ANCA GN progressed to CKD and 12% to ESRD (2). Arulkumaran et al. collated all the pediatric literature on a total of 86 children and found that although the number of children progressing to CKD is similar for both GPA (29%) and MPA (27%), the progression to ESRD is more likely with MPA (35%) than with GPA (5%) (4). Our study could not demonstrate an association between diagnosis of GPA versus MPA and renal outcomes given the small sample size. Overall, renal survival in children (66.7% at longest follow-up for our cohort) appears better than that reported in adults with renal survival estimates of 53.5% at 5 years (20). No child from our cohort died, in stark contrast to the outcomes seen with adult-onset disease.

Our study has limitations, including a small number of children; however, it is one of the largest pediatric cohorts with ANCA GN. We also could not assess the effect of individual therapies on our outcomes of interest given the small sample size and variability in therapy. Given the large time frame over which children were included in the study, and the fact that treatment regimens have changed during this time, our results will need to be validated prospectively, with modern treatment regimens. We also have many children who have transitioned into adulthood and therefore have very little follow-up information after age 18 years. Our study also lacked power to ascertain other factors from within the crescentic/mixed group apart from need for dialysis and degree of proteinuria, which may provide further prognostic information. Perhaps other pathologic factors, such as tubular atrophy and interstitial fibrosis or the degree of CD3+ tubulitis, may help in this regard in future studies (9,23–25).

The clinical utility of this histopathologic classification system and its ability to clearly discriminate kidney outcomes among ANCA GN patients has been demonstrated in adults. This study of 40 children with ANCA GN provides supportive evidence that this classification system is similarly predictive of outcome as in adults and can likely be extended to the pediatric population. This classification will enhance communication to patients and families, clinicians, and researchers. Developing studies to prospectively evaluate treatments by pathologic diagnosis will permit optimization of our treatment strategies and ultimately lead to better evidence for the treatment of this severe disease.

Disclosures

None.