PowerPoint Presentation


Clinical use of the 31-gene expression profile for informing sentinel lymph node biopsies: a prospective, multicenter 
study

Background

›A positive sentinel lymph node (SLN) is an indicator of poor outcomes in cutaneous
melanoma (CM); however, using clinicopathologic factors to select patients results in up to
88% of all patients having a negative SLNB.1

›SLN biopsies (SLNBs) cost up to $15,223 and are associated with complications in 11.3% of
cases.2-3 In order to have a positive impact on survival, 142 SLNBs are needed to find one
patient who will die from their disease.4

›The 31-gene expression profile (GEP) test stratifies patient risk for recurrence, metastasis,
and melanoma death as Class 1A (low-risk), Class 1B/2A (intermediate risk), and Class 2B
(high-risk).

›The 31-GEP is validated to identify patients who have <5% risk of positive SLNB and good
outcomes, indicating that these patients could safely forego the procedure.5-6

Acknowledgments & Disclosures

References

›CNB, KA, CJ, BM, and SJK are employees and shareholders/option holders of Castle Biosciences, Inc.
›JMG is on the speakers bureau at Castle Biosciences. MY has no conflicts to disclose.

Results

Presented at the Winter Clinical 2023 Conferences.

J. Michael Guenther, MD1, Christine N. Bailey, MPH2, Kelli Ahmed, PhD2, Clare Johnson, RN2, Brian Martin, PhD2, Sarah J. Kurley, PhD2, Maki Yamamoto, MD3

1St. Elizabeth Physicians, Edgewood, KY, USA, 2Castle Biosciences, Friendswood, TX, USA, 3University of California-Irvine, Orange, CA, USA

›The 31-GEP result and patient preference play a significant
role in SLNB decisions.

›85% of SLNB decisions were influenced by the 31-GEP result.
83% of SLNB decisions that were discordant with 31-GEP
results were due to patient preference.
›Using the 31-GEP Class 1A result in conjunction with current
guidelines resulted in a 29% decrease in SLNBs compared to
not using the 31-GEP to help guide SLNB decisions.

Conclusions

›Prospectively enrolled patients (n=191) with T1a tumors and at least one high-risk feature, T1b, or T2
tumors were seen by surgical oncologists (89.1%), dermatologists (7.8%), and medical oncologists (3.1%)
from 22 centers.

›Clinicians received 31-GEP results prior to SLNB decisions and were asked which features influenced their
decision whether to perform an SLNB. If the procedure was performed, outcomes were recorded at a
subsequent visit.

›To test the impact of the 31-GEP on SLNB rates, in-study procedure rates were compared to varying
baseline rates using the Exact binomial test.6

›The association between clinicopathological features with SLNB performance was studied using stepwise
selection on a logistic regression model

Descriptor
Class 1A 
(n=138)

Class 1B 
(n=23)

Class 2A 
(n=13)

Class 2B 
(n=19)

All Pts 
(n=193)

P-value

Age, median (range) 65 (25-87) 70 (39-90) 65 (22-84) 69 (50-88) 65 (22-90) 0.118
Breslow thickness, 
median (range)

0.9 (0.2-1.9) 0.8 (0.6-1.85) 1.5 (0.9-1.9) 1.1 (0.4-2) 0.9 (0.2-2) <0.001

Ulceration <0.001
Absent 132 (95.7%) 20 (87.0%) 7 (53.9%) 9 (47.4%) 168 (87.1%)
Present 4 (2.9%) 2 (8.7%) 6 (46.2%) 10 (52.6%) 22 (11.4%)
Unknown 2 (1.5%) 1 (4.4%) 0 (0%) 0 (0%) 3 (1.6%)

Mitotic rate (1/mm2) 0.003
<2 99 (78.0%) 14 (63.6%) 4 (33.3%) 10 (55.6%) 127 (70.9%)
≥2 28 (22.1%) 8 (36.4%) 8 (66.7%) 8 (44.4%) 52 (29.1%)

Physician Specialty 0.037
Surgical Oncologist 125 (90.6%) 17 (73.9%) 13 (100%) 17 (89.5%) 172 (89.1%)
Dermatologist 10 (7.3%) 5 (21.7%) 0 (0%) 0 (0%) 15 (7.8%)
Medical Oncologist 3 (2.2%) 1 (4.4%) 0 (0%) 2 (10.5%) 6 (3.1%)

Figure 1. Physicians use 31-GEP results to guide and implement 
decisions about SLNBs in a risk-aligned manner

Methods

1. Huang, H., et al. Front Oncol 12, 817510 (2022). 2. Agnese et al. Surgery (2003) 3. Moody et al. EJSO 2017. 4.Marchetti et al. JAMA Derm (2021). 5. 

Vetto, J. T. et al. Future Oncology 15, 1207–1217 (2019). 6. Whitman, E.D., et al. JCO-PO 5, 1466-79 (2021).

The 31-GEP result influenced 85% of SLNB decisions. When the 31-GEP result influenced for 
SLNB, 92% were performed, and when the 31-GEP result influenced against SLNB, 70% were not 

performed, indicating that the 31-GEP is used to guide clinical actions.  Of the SLNB decisions 
that were discordant with the 31-GEP result, 83.0% (29/35) were due to patient preference. 

Table 1. Demographic data (n=193)

›Most patients have a negative SLNB
›Identifying patients at low risk of having a positive SLNB can:
›Reduce the number of unnecessary SLNBs
›Reduce SLNB-associated complications
›Reduce healthcare costs

Clinical Implications

Decision Factor Odds Ratio (95% CI)

31-GEP 14.8 (6.1-46.8)*
Patient preference 32.7 (11.5-126.5)*

Tumor-infiltrating lymphocytes 4.7 (1.1-22.5)
Breslow thickness 2.6 (1.2-6.2)

Age 2.1 (0.9-5.1)

Stepwise selection on logistic regression model to identify factors associated with performing an SLNB. *p-
values <0.01; Other variables, while contributing to the model, did not reach the statistical significance 
mark of <0.01 used in this analysis.

Figure 2. SLNB performance rates are significantly reduced when 
incorporating the 31-GEP result into decision-making

Using the 31-GEP in conjunction with current guidelines reduced the number of SLNBs in the entire 
population relative to the baseline control (59.1% vs. 78.0%). In Class 1A patients, the reduction in SLNB 

performance was 29.4%. ** p<0.01

Results

31-GEP Class 1A 1B/2A 2B Total

85%
Influenced
by 31-GEP

SLNB performed
(concordant with
31-GEP influence)
SLNB not performed

SLNB not performed
(concordant with
31-GEP influence)
SLNB performed70%

92%

-29%
-19%

49%

3%

59%

9%

78%

10%

SLNB performance reduction rate, relative to 
contemporary cohort

SLNB positivity rate

SLNB performance rate

P
e
rc

e
n
t,
 %

0

20

40

60

80

100

-20

-40

**

**

Class 1A (low risk), 

n=138

All Pts, n=193

Contemporary 

comparison 

cohort, n=856

Study Design

Patients with CM 
considering 

SLNB 

Physician and patient make 
decision about SLNB after 
receiving 31-GEP results

SLNB

No SLNB

Follow for 
outcomes

Table 2. Factors associated with decision to perform or forego 
SLNB.

31-GEP
testing

Forego SLNB

Perform SLNB

No influence on SLNB

In
fl

u
e

n
c

e
 o

f 
3

1
-G

E
P

 o
n

 S
L

N
B

 
d

e
c

is
io

n
 (

n
u

m
b

e
r 

o
f 

d
e

c
is

io
n

s
)

Scan or click here  for 
more info

https://castlebiosciences.com/research-development/publications/