Patterns of Undertreatment and Overtreatment in Adjuvant Radiotherapy for Early-Stage Endometrial Cancer Based on Molecular Classification

Abstract

The quality improvement study examines the use of risk-adaptive adjuvant radiotherapy in women with non–mismatch repair deficiency endometrial cancer.

Endometrial cancer (EC) is among the few cancers with increasing incidence and mortality over the past 2 decades, prompting substantial effort toward improved disease understanding. International guidelines have been revised to incorporate molecular classification (MC) as part of increasingly personalized management.¹

For early-stage endometrioid EC, adjuvant radiotherapy (RT) recommendations (eFigure 1 in Supplement 1) are supported by comprehensive analyses of genomic profile, locoregional control, and adverse effects in the Post Operative Radiation Therapy in Endometrial Carcinoma (PORTEC)-1 and PORTEC-2 randomized trials²: patients with tumors classified as (1) DNA polymerase-ε mutated (favorable prognosis) should receive no adjuvant RT, (2) no specific molecular profile (intermediate prognosis) should receive vaginal brachytherapy alone, and (3) TP53 mutated (p53abn) (unfavorable prognosis) should receive pelvic external beam RT. As MC grows in practice, we sought to assess the differences in risk-adaptive adjuvant RT guided by traditional clinicopathologic features vs MC as proposed by PORTEC. This study followed the Standards for Quality Improvement Reporting Excellence (SQUIRE) reporting guideline.

Methods

Between 2002 and 2013, 253 women across 21 institutions with International Federation of Gynecology and Obstetrics 2009 stage I endometrioid EC (eTable in Supplement 2) underwent surgical resection and submitted tumor specimens to The Cancer Genome Atlas (TCGA) for MC.³ We excluded 100 patients with mismatch repair deficiency MC (MMRd) from evaluation (eFigure 2 in Supplement 1), as evidence suggests their RT benefit greatly varies with clinicopathologic features.² For the remaining patients, we compared their documented adjuvant RT treatments (preceded TCGA seminal introduction of MC)³ with RT approaches informed by MC, which is strongly prognostic as well as predictive of RT response.² Statistical details are outlined in the eMethods in Supplement 1.

Results

There were 153 women diagnosed with non-MMRd EC (median age, 62 [IQR, 54-71] years) and followed up for a median of 31 (IQR, 20-57) months. Patients with DNA polymerase-ε mutated MC who received any adjuvant RT were considered overtreated (37% [10 of 27]); patients with no specific molecular profile EC who received pelvic external beam RT were considered overtreated (6% [6 of 97]), while those who received no adjuvant RT were considered undertreated (72% [70 of 97]); and patients with p53abn EC who did not receive pelvic external beam RT were considered undertreated (90% [26 of 29]) (Table).

Table. Clinicopathologic Characteristics of Patients With Stage I Endometrioid Endometrial Cancer in The Cancer Genome Atlas.

Characteristic	Adjuvant radiotherapy, No. (%)
	Appropriate treatment (n = 41)	Undertreatment (n = 96)	Overtreatment (n = 16)	P valuea
Age at diagnosis, mean (SD), y	62 (11)	63 (13)	60 (13)	.64
BMI
<30	19 (46)	25 (26)	4 (25)	.11
30 to <40	11 (27)	40 (42)	5 (31)
≥40	8 (20)	28 (29)	5 (31)
Unknown	3 (7)	3 (3)	2 (13)
Menopause status
Pre	3 (7)	8 (8)	3 (19)	.20
Peri	4 (10)	3 (3)	2 (13)
Post	32 (78)	75 (78)	11 (69)
Unknown	2 (5)	10 (10)	0
History of hormone therapy
No	26 (63)	55 (57)	12 (75)	.73
Yes	2 (5)	8 (8)	0
Unknown	13 (32)	33 (34)	4 (25)
Self-reported race and ethnicity
Hispanic	0	1 (1)	1 (6)	.06
Non-Hispanic
American Indian or Alaska Native	0	1 (1)	0
Asian	4 (10)	5 (5)	0
Black	7 (17)	18 (19)	0
Native Hawaiian or Other Pacific Islander	1 (2)	1 (1)	0
White	24 (59)	68 (71)	13 (81)
Unknown	5 (12)	2 (2)	2 (13)
Diagnosis method
Dilation and curettage	8 (20)	23 (24)	2 (13)	.82
Endometrial biopsy	29 (71)	61 (64)	13 (81)
Tumor resection	1 (2)	5 (5)	1 (6)
Other	3 (7)	7 (7)	0
FIGO 2009 stage
IA	22 (54)	76 (79)	8 (50)	.004
IB	15 (37)	12 (13)	5 (31)
I (substage unspecified)	4 (10)	8 (8)	3 (19)
Histologic grade
1	9 (22)	40 (42)	2 (23)	.02
2	18 (43)	24 (25)	4 (32)
3	14 (34)	32 (33)	10 (45)
Molecular classification
POLEmut	17 (41)	0	10 (63)	<.001
NSMP	21 (51)	70 (73)	6 (37)
p53abn	3 (7)	26 (27)	0
Surgical approach
Minimally invasive	17 (41)	38 (40)	4 (25)	.68
Open	22 (54)	55 (57)	12 (75)
Unknown	2 (5)	3 (3)	0
Surgical outcome
R0	31 (76)	68 (71)	12 (75)	.99
R1	0	2 (2)	0
Unknown	10 (24)	26 (27)	4 (25)
Radiotherapy modalityb
None	17 (41)	90 (94)	0	<.001
Vaginal brachytherapy	22 (54)	6 (6)	4 (25)
Pelvic external beam radiotherapy	3 (7)	0	12 (75)

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); FIGO, International Federation of Gynecology and Obstetrics; NSMP, no specific molecular profile; p53abn, TP53 mutated; and POLEmut, DNA polymerase-ε mutated.

^a Continuous and categorial variables were compared across 3 groups using analysis of the variance and Fisher exact test, respectively.

^b Percentages do not total 100% because some patients received both pelvic external beam radiotherapy and vaginal brachytherapy.

Across the entire cohort, 96 patients (63%) were undertreated and 16 (10%) were overtreated. Patients who did not receive any adjuvant RT for p53abn EC, the least favorable molecular class (Figure) with the greatest locoregional benefit from RT in PORTEC,² accounted for 20 of the 96 undertreated patients (21%). Undertreatment was associated with stage IA and low-grade disease, whereas overtreatment was associated with high-grade disease (Table). The distribution of MC also differed by race and ethnicity (Figure). Endometrial cancer classified as p53abn was more common among Black patients (22%) than White (9%; P = .02) or Asian (0%; P = .06) patients.

Figure. Survival and Distribution of Stage I Endometrial Cancer (EC).

Disease-free survival (A) and overall survival (B) for different molecular classes of stage I endometrioid EC in The Cancer Genome Atlas (TCGA). Survival curves were compared using log-rank test. Distribution of stage I endometrioid EC molecular classes in TCGA stratified by race and ethnicity (C). MMRd indicates mismatch repair deficiency; NSMP, no specific molecular profile; p53abn, TP53 mutated; and POLEmut, DNA polymerase-ε mutated.

Discussion

Radiotherapy management of stage I endometrioid EC informed by only traditional clinicopathologic features may predispose a large fraction of patients to receive suboptimal treatment, as discerned by modern precision medicine. Increasing incidence and mortality of EC have been attributed to the emergence of more advanced stage nonendometrioid EC, which disproportionately affects Black women.⁴ Herein, we provide what is, to our knowledge, the first report that even histologically favorable stage I EC consists of patient subgroups with distinctly worse prognosis, also more frequently affecting Black women.

Without refined genomic guidance,² molecularly aggressive disease can be underrecognized, especially if lower substage and histologic grade are present (Table). Broadening access to MC may alleviate outcome disparities for populations that are susceptible to more-concerning disease and yet managed uniformly with other patients due to antiquated risk stratification. The observed racial and ethnic similarity in the frequency of MMRd, a marker of responsiveness to immunotherapy, warrants improved sociodemographic representation⁵ in EC clinical trials as well. While immunotherapies do not play a role in early-stage EC, MMRd identification by proactive MC in the early-stage setting may aid anticipatory management in the recurrent setting where immunotherapies have recently reshaped the treatment landscape.⁶

Supplement 1.

eFigure 1. Choice of Adjuvant Radiotherapy

eFigure 2. Study Schema

eMethods

Supplement 2.

eTable. TCGA Stage I Endometrioid Endometrial Cancer with known molecular classification (POLEmut, p53abn, MMRd, NSMP)

Supplement 3.

Data Sharing Statement