O iginal A icle
Sequence o Con ac X- ay B achy he apy (CXB) and Ex e nal Beam
Radia ion (EBRT) in o gan-p ese ing ea men o small ec al cance
Ngu Wah Than
a
,
b
, D. Ma k P i cha d
a
, Da id M. Hughes
c
, Ca ie A. Duckwo h
a
,
Helen Wong
b
, Muneeb Ul Haq
a
,
b
, Raja am S ipadam
b
, A hu Sun Myin
a
,
b
,
*
a
Depa men o Molecula and Clinical Cance Medicine, Ins i u e o Sys ems, Molecula and In eg a i e Biology, The Uni e si y o Li e pool, L69 3GE, Uni ed Kingdom
b
The Cla e b idge Cance Cen e NHS Founda ion T us , 65 Pemb oke Place, Li e pool L7 8YA, Uni ed Kingdom
c
Depa men o Heal h Da a Science, Ins i u e o Popula ion Heal h, The Uni e si y o Li e pool, L7 3EA, Uni ed Kingdom
ARTICLE INFO
Keywo ds:
Rec al cance
Con ac X- ay b achy he apy
Papillon
Radio he apy
O gan p ese a ion
Wa ch and wai
P opensi y sco e ma ching
In e se p obabili y ea men weigh ing
ABSTRACT
Backg ound and pu pose: Ex e nal Beam Radio he apy (EBRT) ollowed by Con ac X- ay B achy he apy (CXB)
and ice e sa a e iable al e na i es o su ge y o selec ed ec al cance pa ien s who ha e small umou s (≤3
cm). Howe e , he op imal sequence o ea men needs o be es ablished. We compa ed wo app oaches using
P opensi y Sco e (PS) ma ching and in e se p obabili y ea men weigh ing (IPTW) analyses o in es iga e
whe he he sequence o ea men a ec ed pa ien ou comes.
Ma e ials and me hods: This e ospec i e analysis (2008–2019) included pa ien s wi h ec al adenoca cinoma
(cT1-3,N0-1,M0, g ade 1–2, size ≤3 cm) who ecei ed bo h EBRT and CXB, i espec i e o ea men sequence.
PS ma ching and IPTW we e conduc ed o balance co a ia e s anda dised mean di e ences be ween g oups.
Oncological ou comes and a e o pos - ea men ec al bleeding we e assessed.
Resul s: Following PS ma ching and IPTW analyses om 251 eligible pa ien s; 103 s a ing wi h EBRT (median
ollow-up: 37 [IQR:18–56] mon hs) and 148 wi h CXB (median ollow-up: 32 [IQR:16–54] mon hs, a signi ican
imp o emen in 3-yea o e all su i al (77% s 85%, p =0.02, [HR:0.58 (95% CI:0.37–0.91)]) and a highe isk
o pos - ea men ec al bleeding (g ade 1 (26%) and g ade 2 (6%)) we e ound in pa ien s who s a ed wi h CXB
(p =0.08). No signi ican di e ences we e obse ed in local eg ow h (18% s 12%, p =0.47), dis an elapse
(10% s 6%, p =0.53), 3-yea o gan p ese a ion a es (70% s 75%, p =0.20, [HR:0.66 (95% CI: 0.35–1.26)]),
o disease- ee su i al (78% s 82%, p =0.17, [HR: 0.47 (95% CI: 0.16–1.38)])
Conclusion: In pa ien s wi h ec al cance (≤3 cm), commencing wi h CXB a he han EBRT, was associa ed wi h
imp o ed o e all su i al, bu had a highe isk o G1/2 ec al bleeding. No s a is ically signi ican di e ences
we e obse ed in o he oncological ou comes.
In oduc ion
Colo ec al cance is he hi d mos common cance and he second
mos common cause o cance dea hs wo ldwide [1]. O gan-p ese ing
ea men op ions o ec al cance ha e been used as iable al e na i es
o su ge y o many yea s, in o de o a oid undesi able su gical isks
and hei impac s on pa ien s’quali y o li e. These ea men s a e used
pa icula ly in pa ien s who a e ail o who ha e signi ican co-
mo bidi ies and in hose who a e s oma a e se [2–12]. The combina-
ion o Con ac X- ay B achy he apy (CXB) and Ex e nal Beam
Radio he apy (EBRT) as an o gan p ese a ion app oach has been
shown o ha e p o en bene i s in sus aining local umou con ol and
main aining long- e m a ou able bowel unc ion [6–8,13–15].
Published analyses by Papillon e al., Schild e al., and Aumock e al.
o pa ien ou comes a e s a ing wi h EBRT ollowed by CXB ha e
e ealed 5-yea local umou con ol a es anging om 71%-92%,
while he disease- ee su i al was 60–76% o T2-3, small (≤3cm)
ec al cance cases [16–18].
Howe e , ew s udies ha e in es iga ed whe he s a ing wi h CXB
ins ead o EBRT esul s in any imp o emen in local umou con ol o
* Co esponding au ho a : The Cla e b idge Cance Cen e NHS Founda ion T us , 65 Pemb oke Place, Li e pool L7 8YA, Uni ed Kingdom.
E-mail add esses: [email p o ec ed] (N.W. Than), [email p o ec ed] (D.M. P i cha d), [email p o ec ed] (D.M. Hughes), C.A.
[email p o ec ed] (C.A. Duckwo h), [email p o ec ed] (H. Wong), [email p o ec ed] (M.U. Haq), [email p o ec ed]
(R. S ipadam), [email p o ec ed] (A.S. Myin ).
Con en s lis s a ailable a ScienceDi ec
Radio he apy and Oncology
jou nal homepage: www. heg eenjou nal.com
h ps://doi.o g/10.1016/j. adonc.2024.110465
Recei ed 15 Feb ua y 2024; Recei ed in e ised o m 18 July 2024; Accep ed 25 July 2024
Radio he apy and Oncology 199 (2024) 110465
A ailable online 27 July 2024
0167-8140/© 2024 The Au ho (s). Published by Else ie B.V. This is an open access a icle unde he CC BY license ( h p://c ea i ecommons.o g/licenses/by/4.0/ ).
educed local ecu ence, pa icula ly o small ec al cance s (≤3 cm).
The s udy by Dhadda e al., in which pa ien s ecei ed CXB p io o
EBRT, in he se ing o pos -ope a i e ea men a e local umou
excision, showed a lowe isk o local umou ecu ence (4–5% s 15%)
compa ed o pa ien s who ecei ed EBRT i s [19]. Beneze y e al.
simila ly ound ha a ea men app oach in ol ing up on CXB o
umou s ≤3 cm esul ed in a mo e a ou able ini ial clinical comple e
esponse (88% s 32%) compa ed o EBRT i s [20]. The OPERA
andomised phase 3 con olled clinical ial also demons a ed an
excellen 3-yea o gan p ese a ion a e in he CXB i s a m (97%)
compa ed wi h he EBRT i s a m (68%) [15]. Howe e , in his ial,
CXB was only used i s o smalle umou s (≤3 cm), whe eas EBRT was
used ini ially o la ge umou s (>3 cm). Hence, he esul s o OPERA
a e no ully able o in o m whe he one ea men sequence was p e -
e able o he o he in pa ien s who had small umou s.
Gaps in he e idence he e o e emain in add essing he op imal
sequence o ea men , whe he o s a wi h CXB i s o EBRT, pa ic-
ula ly o small ec al cance s (≤3 cm). Recen ly, a new guideline om
he F ench Na ional Socie y o Gas oen e ology (SNFGE) has ecom-
mended CXB as he ini ial ea men o cT2N0M0 ec al cance wi h a
umou diame e o ≤3 cm [21]. Howe e , a ious sequences o his
o gan-p ese ing ea men a e s ill being used in e na ionally.
We he e o e aimed o compa e he oncological ou comes be ween
hese wo ea men app oaches. We used p opensi y sco e ma ching
and in e se p obabili y ea men weigh ing (IPTW) analyses o op i-
mise he e ec i eness o he ga he ed obse a ional da a and ensu e a
balanced dis ibu ion o baseline cha ac e is ics be ween he wo
ea men g oups.
Ma e ials and me hods
Pa ien selec ion
Ins i u ional audi app o al was ob ained o a e ospec i e e iew
o all sequen ial pa ien s who ecei ed CXB ea men a Cla e b idge
Cance Cen e (CCC), Me seyside, UK om 2008 o 2019. Da a om
pa ien s who had unde gone bo h EBRT and CXB wi h cu a i e in en ,
ega dless o he ea men sequence, we e collec ed om he da abase
sys em including demog aphics, umou size and s age, ea men de-
ails and ou come da a. Only pa ien s who had well o mode a ely
di e en ia ed localised ec al adenoca cinomas (cT1-3, cN0-1, cM0),
small umou diame e s (≤3 cm), and no ea u es o ex amu al enous
in asion we e included in he cu en analysis. The umou diame e
was de e mined om p e- ea men MRI scan epo s, digi al ec al ex-
amina ions (DRE) pe o med by he oncologis o colo ec al su geon,
and he CXB applica o size ha was used du ing he i s ea men
session.
Pa ien managemen
Pa ien s we e e e ed o CCC by local Mul idisciplina y Teams
(MDTs) a colo ec al cance cen es h oughou he UK. In some cases,
pa ien s had al eady ecei ed EBRT locally and we e e e ed o
conside a ion o CXB because hey had no had a comple e esponse o
his ea men . O he pa ien s we e e e ed de no o p io o ecei ing
any he apy and in some o hese, i was decided o adminis e CXB i s
ollowed by EBRT. The decision o begin ea men wi h ei he (chemo)
adia ion o e e pa ien s o conside a ion o s a ing wi h CXB was
based on he p e e ence o he local MDT, as he e a e only ou CXB
ea men cen es in he UK.
EBRT was adminis e ed ei he as chemo adia ion (45–50.4 Gy/25
#/35 days), long-cou se (40–45 Gy/20#/28 days), o sho -cou se (25
Gy/5#/5 days) egimens. The decision o choose a pa icula egimen
was based on he pa ien ’s pe o mance s a us and como bidi ies.
Con o mal 3-dimen ional adia ion was he p ima y echnique be o e
2013 wi h a shi o In ensi y Modula ed Radia ion The apy (IMRT)
he ea e . Radia ion olume was ypically limi ed o ea he meso-
ec um and gene ally did no go abo e S2/3. No adju an chemo he apy
was ou inely adminis e ed a e adio he apy ea men .
CXB was pe o med on an ou pa ien basis using he Papillon-50
machine (50kVp X- ays (HVL 0.64 Al, 2.7 mA), A iane, Al e on, UK),
20-30 Gy pe ac ion deli e ed 2 weeks apa , h ough a ec al ea -
men applica o (size 30, 25, o 22 mm) a a ocal sou ce su ace dis ance
o 29, 32, o 38 mm espec i ely. A o al dose o 90–110 Gy ( ec al
mucosal su ace dose) was deli e ed in 3–4 ac ions o e 4–6 weeks.
Pa ien s who achie ed a clinical/nea comple e esponse 8–12 weeks
a e comple ion o all ea men s we e assessed egula ly wi h a DRE,
lexible sigmoidoscopy/ igid ec oscopy al e na ely a hei local cen e
by he colo ec al su geon and by he adia ion oncologis a CCC excep
o pa ien s who we e unable o a end egula ollow-ups a he CCC,
and MRI scans a 12-week in e als a hei local cen e in he i s 2
yea s and e e y 6 mon hs in he 3 d yea . Usually, only endoscopic
examina ion was done in he 4 h and 5 h yea s. Pa ien s who did no
achie e a comple e esponse we e e e ed o sal age su ge y i hey
we e i enough o his.
Ou come measu es
The p ima y ou come measu es we e o e all su i al (OS) and
disease- ee su i al (DFS). Disease- ee su i al was calcula ed om
he da e o he las adio he apy ea men o he da e o loco egional
ecu ence a e R0/1 esec ion o he p ima y umou , non-sal ageable
local eg ow h/R2 sal age esec ion, he occu ence o second p ima y,
dis an me as asis, o las ollow-up. O e all su i al was de ined as he
pe iod om he las adio he apy ea men o he da e o he las da a
e iew o dea h om any cause. Seconda y ou come measu es included
he a e o ini ial clinical Comple e Response (cCR), local eg ow h,
nodal/ egional elapse, dis an elapse, o gan p ese a ion a e, and
adia ion oxici ies, speci ically pos - ea men ec al bleeding.
S a is ical analysis
We u ilised wo p opensi y sco e me hods o minimise he s an-
da dized mean di e ences in co a ia es, he eby achie ing be e -
balanced coho s in his e ospec i e single-cen e analysis. Typically,
he p opensi y sco e ma ching is calcula ed h ough mul i a ia e logis ic
eg ession o ming a ma ched coho . Howe e , ma ching o en esul s
in a educed sample size, leading o biased es ima ion in su i al anal-
ysis in small samples wi h low e en a es. To o e come his, we also
employed IPTW analysis which main ains in o ma ion om mos pa-
ien s by c ea ing a syn he ic sample, in which he dis ibu ion o
measu ed baseline co a ia es is independen o ea men assignmen
[22,23].
Fi s ly, he associa ion be ween adio he apy egimen and indi idual
ime- o-e en ou come (OS and DFS) was examined using su i al
model in an unadjus ed analysis. Secondly, conside ing he po en ial
in luence o a ious baseline co a ia es, he second se o su i al
models adjus ed o hese co a ia es o each ou come. The a iables o
age, gende , i ness o su ge y, pe o mance s a us, umou (T) s age,
nodal (N) s age, EBRT egimen, CXB o al dose, and o e all ea men
ime we e conside ed possible con ounde s. The o e all ea men ime
was calcula ed om he s a da e o he end da e o all adio he apy
ea men s, including any gaps be ween he apies. This du a ion was
hen di ided in o wo ca ego ies. S anda d ea men ime was de ined
as ≤12 weeks (EBRT =1week, maximum gap ime =4 weeks, o al CXB
ea men ime =4–6 weeks) o pa ien s who ecei ed sho -cou se
ex e nal beam adia ion o ≤25 weeks (EBRT =5 weeks, maximum
gap ime =14 weeks, o al CXB ea men ime =4–6 weeks) o hose
who ecei ed long-cou se (chemo) adia ion. Uncon en ional ea men
ime was designa ed as ≥12 weeks o pa ien s who ecei ed sho -
cou se ex e nal beam adia ion o ≥25 weeks o hose who ecei ed
long-cou se (chemo) adia ion.
N.W. Than e al. Radio he apy and Oncology 199 (2024) 110465
2
A e ha , we adop ed wo p opensi y sco e weigh ing me hods o
minimise he in luence o con ounding a iables in ou e ospec i e,
obse a ional s udy. The p opensi y sco e indica es he p obabili y o an
indi idual ecei ing a speci ic ea men , as de e mined by p ede ined
co a ia es. This p obabili y is calcula ed using a logis ic eg ession
model in which ea men is ea ed as a bina y dependen a iable, and
a designa ed se o co a ia es se es as independen a iables o ensu e
balance. The p opensi y sco es we e used in wo ways. Fi s , we pe -
o med p opensi y sco e ma ching using he nea es -neighbou me hod
wi h a callipe o 0.25. Then, IPTW was calcula ed based on he p o-
pensi y sco es o weigh he ull coho . The balance o co a ia es be o e
and a e p opensi y ma ching and IPTW was assessed by examining he
s anda dised mean di e ence (SMD) be ween he g oups. An SMD o
g ea e han 0.1 is conside ed a signi ican imbalance. Su i al analysis
me hods we e used o e alua e he wo ime- o-e en ou comes. Visual
inspec ion o su i al di e ences using Kaplan-Meie cu es and s a-
is ical assessmen s using Log- ank es s we e pe o med. The associa-
ion be ween ea men ype and su i al isks was analysed using Cox
p opo ional haza ds models, o each ime- o-e en ou come, he p o-
po ional haza ds assump ion equi ed by he Cox model was assessed
by Schoen eld esiduals. P- alue <0.05 was conside ed s a is ically
signi ican . S a is ical analyses we e pe o med in R 4.3.0.
Resul s
A o al o 251 eligible pa ien s, who ecei ed bo h EBRT and CXB,
ega dless o he sequence, wi h cu a i e in en we e included in he
s udy. One hund ed and h ee pa ien s ecei ed EBRT and 148 had CXB
as hei ini ial ea men . The median ollow-up was 37 [IQR:18–56]
mon hs o he EBRT- i s g oup and 32 [IQR:16–54] mon hs o he
CXB- i s g oup.
The wo g oups we e easonably ma ched in e ms o age and gende
p o iles, pe o mance s a us and i ness o su ge y. Howe e , signi ican
co a ia e imbalances in e ms o pa ien s’ umou (T) s age, nodal (N)
s age, EBRT egimens, and p olonged o e all ea men ime we e
e iden be ween he wo g oups (Table 1). Pa ien s in he EBRT i s
g oup had signi ican ly mo e ad anced umou s ages and we e also
mo e likely o ha e ecei ed concu en chemo he apy as pa o hei
EBRT egimen. P opensi y sco e ma ching and IPTW we e he e o e
pe o med o conside hese di e ences and o allow a mo e app op ia e
s a is ical analysis. Following PS ma ching and IPTW, he e was a good
balance ac oss all co a ia es, wi h no s a is ically signi ican di e ences
emaining apa om sligh de ia ions o SMD obse ed o he dis ance
om he anal e ge in he p opensi y-ma ched model, and o i ness o
su ge y and EBRT egimens in he IPTW model (Supplemen a y Fig. 1A
and 1B). The demog aphic da a o all coho s a e summa ised in
Table 1.
In he unadjus ed analysis, he a es o ini ial cCR 79% s 88%, (p =
0.07), local eg ow h 19% s 15%, (p =0.81), nodal/ egional elapse
4% s 2.7%, (p =0.60), and dis an elapse 13% s 8%, (p =0.24) we e
obse ed in he EBRT- i s and CXB- i s g oups, espec i ely (Table 2).
The 3-yea o gan p ese a ion a es we e 69% (95%CI: 55–78) and 73%
(95%CI: 63–79) (p =0.42, HR: 0.82 [95%CI: 0.51–1.32]) (Fig. 2A) wi h
3-yea disease- ee su i al a es o 78% (95%CI: 72–82) and 80% (95%
Table 1
Baseline cha ac e is ics o unadjus ed, p opensi y-sco e ma ched, and IPTW coho s.
Cha ac e is ics Unadjus ed P opensi y sco e-ma ched IPTW
EBRT i s
n=103(%)
CXB i s
n=148(%)
P alue EBRT i s
n=72(%)
CXB i s
n=72(%)
P alue EBRT i s
n=290(%)
CXB i s
n=243(%)
P alue
Mean age 71.9 ±10.8 73.7 ±11.2 0.22 71.6 ±11.0 69.3 ±11.8 0.24 72.1 ±9.3 72.9 ±11.0 0.52
Gende Male
Female
75(73)
28(27)
101(68)
47(32)
0.44 54(75)
18(25)
50(69)
22(31)
0.58 203(71)
87(29)
165(68)
78(32)
0.81
Fi ness o su ge y Fi
Un i
75(73)
28(27)
101(68)
47(32)
0.44 32(44)
40(56)
37(51)
35(49)
0.51 154(53)
136(47)
121.5(50)
121.5(50)
0.39
WHO pe o mance s a us 1–2
2–3
63(61)
40(39)
78(53)
70(47)
0.18 46(64)
26(36)
47(65)
25(35)
1.00 180(62)
110(38)
124(51)
119(49)
0.82
T s age T1
T2
T3
3(3)
51(49)
49(48)
15(10)
104(70)
29(20)
<0.001 3(4)
44(61)
25(35)
4(6)
46(64)
22(31)
0.83 14(5)
189(65)
87(30)
17(7)
158(65)
68(28)
0.82
N s age N0
N1
67(65)
36(35)
128(87)
20(13)
<0.001 56(78)
16(22)
56(78)
16(22)
1.0 235(81)
55(19)
190(78)
53(22)
0.70
Dis ance om anal e ge <6cm
≥6–12 cm
65(63)
38(37)
105(71)
43(29)
0.19 45(63)
27(37)
56(78)
16(22)
0.69 220(76)
70(24)
175(72)
68(28)
0.53
EBRT egimen Sho cou se
Long cou se
Chemo adia ion
21(20)
21(20)
61(60)
82(55)
13(9)
53(36)
<0.001 20(28)
15(21)
37(51)
22(31)
9(12)
41(57)
0.41 114(45)
33(13)
106(42)
111(45)
32(13)
104(42)
0.96
CXB dose 80-90gy
110gy
78 (76)
25(24)
111 (75)
37(25)
1.0 57(79)
15(21)
52(72)
20(28)
0.44 218(75)
72(25)
180(74)
63(26)
0.77
O e all ea men ime S anda d
Uncon en ional
67(65)
36(35)
123(83)
25(17)
0.002 56(78)
16(22)
56(78)
16(22)
1.00 232(80)
58(20)
190(78)
53(22)
0.71
Table 2
Oncological ou comes o di e en analyses.
Ou come Unadjus ed P opensi y sco e-ma ched IPTW
EBRT i s
n=103(%)
CXB i s
n=148(%)
P alue EBRT i s
n=72(%)
CXB i s
n=72(%)
P alue EBRT i s
n=290(%)
CXB i s
n=243(%)
P alue
Ini ial esponse 81(79) 130(88) 0.07 46 (78) 54(85) 0.39 249(86) 214(88) 0.57
Local eg ow h 15(19) 20(15) 0.81 10 (18) 7 (12) 0.47 55(19) 24(10) 0.20
Nodal/ egional elapse 4(4) 4(2.7) 0.60 2(3) 1(1.4) 1.00 9(3) 8(3) 1.00
Dis an elapse 13(13) 12(8) 0.24 7(10) 4(6) 0.53 23(8) 17(7) 0.75
Absolu e o gan p ese a ion a e 71(69) 111(75) 0.29 49 (68) 56 (78) 0.26 212 (73) 197(81) 0.26
Rec al bleeding
G1
G2
18(18)
15(15)
3(3)
48(32)
38 (26)
10(6)
0.01 13 (18) 23 (32) 0.08 38(13) 90(37) 0.004
N.W. Than e al. Radio he apy and Oncology 199 (2024) 110465
3
CI: 75–85). Cox p opo ional haza d model o disease- ee su i al
showed no signi ican di e ence be ween he wo egimens in bo h
uni a iable (p =0.68, HR: 0.87 [95%CI: 0.45–1.67]) and mul i a iable
analyses (p =0.66, HR: 0.84 [95%CI: 0.39–1.82]) (Fig. 3A). The 3-yea
o e all su i al a es we e 78% (95%CI: 73–84) and 79% (95% CI:
74–84) and no s a is ical di e ence in haza d a io was obse ed be-
ween he wo g oups in uni a iable analysis (p =0.44, HR: 0.88 [95%
CI: 0.63–1.22]). Howe e , he mul i a iable analysis indica ed a
Fig. 1. Flow cha illus a ing o e all clinical ou comes o he en i e s udy g oup.
N.W. Than e al. Radio he apy and Oncology 199 (2024) 110465
4
signi ican di e ence, sugges ing ha pa ien s ecei ing CXB i s we e
mo e likely o su i e (p =0.03, HR: 0.65 [95%CI: 0.43–0.97])
(Fig. 4A). Pos - ea men G1-2 sel -limi ing ec al bleeding occu ed in
18% o EBRT- i s pa ien s and 32% o CXB- i s pa ien s (p =0.01). The
o e all clinical ou comes o he en i e s udy g oup a e illus a ed in
Fig. 1.
In he p opensi y-ma ched coho , he ini ial cCR a e was 78% s
85%, (p =0.39) wi h he local eg ow h a e o 18% s 12%, (p =0.47),
nodal/ egional elapse a e o 3% s 1.4%, (p =1.00), and dis an
elapse a e o 10% s 6%, (p =0.53) o he EBRT- i s and CXB- i s
egimens espec i ely (Table 2). The 3-yea o gan p ese a ion a es
we e 70% (95%CI: 55–80) and 75% (95%CI: 72–85), (p =0.20, HR: 0.66
[95%CI: 0.35–1.26]) (Fig. 2B) wi h 3-yea disease- ee su i al a es o
78% (95%CI: 72–90) and 82% (95%CI: 78–97). The Cox model o
disease- ee su i al indica ed no signi ican di e ence be ween g oups
in bo h uni a iable (p =0.17, HR: 0.47 [95%CI: 0.16–1.38]) and
mul i a iable analyses (p =0.23, HR: 0.50 [95%CI: 0.16–1.55])
(Fig. 3B). The 3-yea o e all su i al a es we e 77% (95%CI: 73–83)
and 85% (95%CI: 80–95). Signi ican di e ences in haza d a ios be-
ween he wo g oups we e obse ed in bo h uni a iable (p =0.02, HR:
0.58 [95%CI: 0.37–0.91]) and mul i a iable (p =0.001, HR: 0.44 [95%
CI: 0.28–0.73]) analyses, indica ing ha pa ien s ecei ing CXB i s
we e mo e likely o su i e (Fig. 4B). Pos - ea men ec al bleeding
occu ed in 18 % o EBRT- i s pa ien s and 32 % o CXB- i s pa ien s (p
=0.08).
The IPTW analysis e ealed he ollowing ou comes o EBRT- i s
and CXB- i s ea men s, espec i ely: ini ial cCR a es (86% s. 88%,
p=0.57), local eg ow h a es (19% s. 10%, p =0.20), simila nodal/
egional a es (bo h 3%, p =1.0), and dis an elapse a es (8% s. 7%, p
=0.75) as shown in Table 2. The 3-yea o gan p ese a ion a es we e
73% (95%CI: 55–80) and 80% (95%CI: 72–90), (p =0.27, HR: 0.47
[95%CI: 0.35–1.10]), wi h 3-yea disease- ee su i al o 87% (95 %CI:
76–92) and 88% (95%CI: 78–95), and o e all su i al o 78% (95%CI:
65–85) and 78% (95%CI: 70–82). Cox p opo ional haza d a ios o
Fig. 2A. Kaplan-Meie cu e compa ing o gan-p ese a ion a es be ween wo
sequences o ea men in unadjus ed model.
Fig. 2B. Kaplan-Meie cu e compa ing o gan-p ese a ion a es be ween wo
sequences o ea men in p opensi y-ma ched model.
N.W. Than e al. Radio he apy and Oncology 199 (2024) 110465
5
disease- ee su i al indica ed no signi ican di e ence be ween he wo
egimens in bo h uni a iable (p =0.20, HR: 0.46 [95%CI: 0.16–1.51])
and mul i a iable analyses (p =0.95, HR: 1.03 [95%CI: 0.46–2.29]).
Despi e no s a is ical signi icance (p =0.06, HR: 0.70 [95%CI:
0.48–1.03]) being obse ed in uni a iable analysis, he mul i a iable
analysis demons a ed a signi ican imp o emen in o e all su i al
wi h CXB- i s ea men (p =0.005, HR: 0.58 [95%CI: 0.40–0.85]). A
highe isk o pos - ea men ec al bleeding was also seen in he CXB-
i s a m (p =0.004) (Supplemen a y Table 1).
Discussion
Al hough commencing o gan-p ese ing ea men wi h CXB ol-
lowed by EBRT o small ec al umou s (≤3 cm) has p e iously sug-
ges ed ha his he apy sequence esul s in imp o ed local umou
con ol a es [15,20], he op imal ea men selec ion o small ec al
cance s has no ye been es ablished. I is he e o e impe a i e o
in es iga e whe he deli e ing CXB i s has any bene i s o e he
adminis a ion o EBRT as he ini ial ea men , pa icula ly in e ms o
long- e m clinical ou comes.
The pa ien and umou cha ac e is ics p esen ed in his s udy
closely esembled hose ou lined in he OPERA ial and Beneze y e al.
s udies o he CXB i s g oup, excep o a somewha lowe pe o -
mance s a us (WHO 2–3) (44% s. 9–10%) and a highe isk o su ge y
(57%). In addi ion, nea ly hal o he pa ien s (41%) in ou s udy un-
de wen a sho cou se o EBRT, whe eas almos all pa ien s in he o he
s udies ecei ed a long cou se o chemo adia ion [15,20].
In he cu en s udy, pa ien s who unde wen CXB as hei ini ial
ea men exhibi ed an ini ial esponse a e o 85–88% which is e y
simila o he indings p esen ed in Beneze y e al. (88%). The a es o
local eg ow h (10–15%) and dis an elapse (6–8%) we e also compa-
able o hose s udies. The 3-yea o gan p ese a ion a e was 73–80%,
whe eas he OPERA ial epo ed 97% o gan p ese a ion a 3 yea s
[15]. A highe isk o pos - ea men occasional sel -limi ing g ade 1–2
Fig. 3A. Kaplan-Meie cu e compa ing disease- ee su i al be ween wo se-
quences o ea men in unadjus ed model.
Fig. 3B. Kaplan-Meie cu e compa ing disease- ee su i al be ween wo se-
quences o ea men in p opensi y-ma ched model.
N.W. Than e al. Radio he apy and Oncology 199 (2024) 110465
6
ec al bleeding was also seen in he CXB i s g oup, consis en wi h he
indings o he OPERA s udy. This migh be ela ed o a highe incidence
o pa ien s who we e on an icoagulan he apy in he CXB- i s g oup
(23% s 14%).
Fo he pa ien s in his coho who s a ed wi h EBRT, he ini ial cCR
a e was 78–86%, he local eg ow h a e was 18–19%, and he dis an
elapse a e was 8–13%. The 3-yea o gan p ese a ion a e was 69–73%
wi h a 3-yea disease- ee su i al a e o 78–87%. These indings a e
compa able o he ou comes in he published s udies by Papillon e al.,
Schild e al., and Aumock e al. [16–18].
We ound a signi ican imp o emen in o e all su i al among pa-
ien s who unde wen CXB as hei ini ial ea men (79–85% a 3 yea s),
compa ed o pa ien s who ecei ed EBRT i s (77–78% a 3 yea s),
despi e no signi ican di e ence being obse ed in disease- ee su i al
(80–88% s 78–87% a 3 yea s). One possible explana ion could be ha ,
despi e no eaching s a is ical signi icance, he CXB- i s a m exhibi ed
lowe a es o local eg ow h and dis an elapse, along wi h a highe
a e o o gan p ese a ion, in bo h unadjus ed and p opensi y-ma ched
coho s. Addi ionally, some pa ien s who unde wen sal age su ge y
a e ailing o gan-p ese ing ea men we e only ollowed up a hei
local cen es and no longe had appoin men s a CCC. Consequen ly, he
disease- ee su i al endpoin , ma ked by he las ollow-up da e
eco ded in he CCC da abase, appea ed o be sho e . Howe e , he
s a us o pa ien s as being ali e/deceased was con inuously upda ed in
he da abase sys em and is he e o e mo e accu a e. These ac o s could
ha e con ibu ed o he obse ed sho e disease- ee su i al pe iod
while po en ially ex ending he o e all su i al.
Despi e conduc ing p opensi y-ma ching analyses, ou s udy s ill
e ains se e al po en ial con ounde s which may ha e po en ially
a ec ed he esul s. The majo limi a ion is he inhe en na u e o a non-
andomized e ospec i e obse a ional s udy, which canno comple ely
add ess unobse ed di e ences be ween he g oups.
The o gan p ese a ion a e ha we ound may ha e been in luenced
by a ia ions in sal age su ge y o esidual/local eg ow h be ween he
Fig. 4A. Kaplan-Meie cu e compa ing o e all su i al be ween wo se-
quences o ea men in unadjus ed model.
Fig. 4B. Kaplan-Meie cu e compa ing o e all su i al be ween wo se-
quences o ea men in p opensi y-ma ched model.
N.W. Than e al. Radio he apy and Oncology 199 (2024) 110465
7
wo g oups. Sal age local excision (TEMS o TAMIS) was pe o med in 5
ou o 37 pa ien s in he EBRT- i s g oup, whe eas only 3 ou o 38
pa ien s in he CXB- i s g oup unde wen local excision (Fig. 1). Addi-
ionally, mo e han hal o he pa ien s om bo h coho s we e ail and
we e a high isk o su ge y, so un o una ely, hey we e no i enough
o unde go sal age su ge y o hei esidual/ ecu en diseases. This
may po en ially ha e con ibu ed o he obse ed educ ion in disease-
ee su i al despi e he mo e p olonged o e all su i al.
I is also impo an o no e he e was he e ogenei y in he ea men
egimens ha we e used wi hin he CXB- i s g oup: 55 pa ien s ecei ed
a comple e cou se o CXB ollowed by EBRT whe eas 93 pa ien s un-
de wen a spli cou se o CXB and EBRT (1–3 ac ions o CXB ollowed
by EBRT and hen ano he 1–2 ac ions o CXB o any esidual disease).
Pa ien s who ecei ed a spli cou se o ea men ended o exhibi a
mo e limi ed ini ial esponse o CXB ea men , leading o he admin-
is a ion o his speci ic egimen. This a iabili y may ha e in oduced
complexi ies in assessing he ue ou comes o his ea men app oach.
Ano he impo an ca ea o ou s udy ela es o he ini ial me hod o
pa ien e e al. I is known ha app oxima ely 30% o ec al cance
pa ien s achie e a clinical comple e esponse ollowing ini ial neo-
adju an (chemo) adia ion [24]. Typically, hese pa ien s a e moni o ed
h ough ac i e su eillance and a e no e e ed o addi ional ea men
unless hey expe ience local/dis an disease elapse. I is he e o e
p obable ha some pa ien s who would o he wise ha e been eligible o
his s udy, expe ienced a sus ained comple e clinical esponse ollowing
EBRT a hei local ea men cen e and we e he e o e no e e ed o
CCC o u he e alua ion.
Conclusions
Despi e hese limi a ions, ou s udy sugges s ha commencing
ea men o small ec al cance (≤3 cm) wi h CXB, as opposed o EBRT,
appea s o be associa ed wi h inc eased o e all su i al, despi e a highe
incidence o occasional sel -limi ing g ade 1–2 ec al bleeding. These
indings align, in pa , wi h o he eme ging e idence om ecen
s udies. We he e o e ecommend ha adequa ely powe ed p ospec i e
clinical ials a e pe o med in o de o de ini i ely es ablish whe he
CXB- i s p o ides bene i s o e EBRT- i s in small ec al cance s.
CRediT au ho ship con ibu ion s a emen
Ngu Wah Than: W i ing –o iginal d a , Visualiza ion, So wa e,
Me hodology, In es iga ion, Fo mal analysis, Da a cu a ion, Concep u-
aliza ion. D. Ma k P i cha d: W i ing – e iew &edi ing, Valida ion,
Supe ision, P ojec adminis a ion, Me hodology, Funding acquisi ion,
Da a cu a ion, Concep ualiza ion. Da id M. Hughes: Valida ion, Su-
pe ision, So wa e, Me hodology, Da a cu a ion, Concep ualiza ion.
Ca ie A. Duckwo h: Supe ision, Me hodology, Concep ualiza ion.
Helen Wong: Resou ces, Da a cu a ion. Muneeb Ul Haq: Da a cu a ion.
Raja am S ipadam: Resou ces, Da a cu a ion. A hu Sun Myin :
W i ing – e iew &edi ing, Supe ision, Resou ces, P ojec adminis-
a ion, Me hodology, Funding acquisi ion, Da a cu a ion,
Concep ualiza ion.
Decla a ion o compe ing in e es
The au ho s decla e ha hey ha e no known compe ing inancial
in e es s o pe sonal ela ionships ha could ha e appea ed o in luence
he wo k epo ed in his pape .
Acknowledgmen
This p ojec was unded by he Eu opean Union’s Ho izon 2020
esea ch and inno a ion p og amme, Ma ie Skłodowska-Cu ie g an
ag eemen numbe 857894—CAST.
Appendix A. Supplemen a y ma e ial
Supplemen a y da a o his a icle can be ound online a h ps://doi.
o g/10.1016/j. adonc.2024.110465.
Re e ences
[1] Cance I.A. .R.o. 10_8_9-Colo ec um- ac -shee ; 2020. <h ps://gco.ia c. / oday/
da a/ ac shee s/cance s/10_8_9-Colo ec um- ac -shee .pd />[accessed 12
No embe 2023].
[2] Hab -Gama A, S˜
ao Juli˜
ao GP, Vaila i BB, e al. O gan p ese a ion in cT2N0 ec al
cance a e neoadju an chemo adia ion he apy: he impac o adia ion he apy
dose-escala ion and consolida ion chemo he apy. Ann Su g 2019;269:102–7.
h ps://doi.o g/10.1097/SLA.0000000000002447.
[3] Renehan AG, Malcomson L, Emsley R, e al. Wa ch-and-wai app oach e sus
su gical esec ion a e chemo adio he apy o pa ien s wi h ec al cance ( he
OnCoRe p ojec ): a p opensi y-sco e ma ched coho analysis. Lance Oncol 2016;
17:174–83. h ps://doi.o g/10.1016/S1470-2045(15)00467-2.
[4] an de Valk MJM, Hilling DE, Bas iaanne E, e al. Long- e m ou comes o clinical
comple e esponde s a e neoadju an ea men o ec al cance in he
In e na ional Wa ch &Wai Da abase (IWWD): an in e na ional mul icen e
egis y s udy. Lance 2018;391:2537–45. h ps://doi.o g/10.1016/S0140-6736
(18)31078-X.
[5] Papillon J. P esen s a us o adia ion he apy in he conse a i e managemen o
ec al cance . Radio he Oncol 1990;17:275–83. h ps://doi.o g/10.1016/0167-
8140(90)90001-d.
[6] Ge a d JP, Chape O, Nemoz C, e al. Imp o ed sphinc e p ese a ion in low ec al
cance wi h high-dose p eope a i e adio he apy: he Lyon R96–02 andomized
ial. J Clin Oncol 2004;22:2404–9. h ps://doi.o g/10.1200/JCO.2004.08.170.
[7] Appel AL, Pløen J, Ha ling H, e al. High-dose chemo adio he apy and wa ch ul
wai ing o dis al ec al cance : a p ospec i e obse a ional s udy. Lance Oncol
2015;16:919–27. h ps://doi.o g/10.1016/S1470-2045(15)00120-5.
[8] Sun Myin A, Smi h FM, Gollins S, e al. Dose escala ion using con ac X- ay
b achy he apy a e ex e nal beam adio he apy as nonsu gical ea men op ion
o ec al cance : ou comes om a single-cen e expe ience. In J Radia Oncol Biol
Phys 2018;100:565–73. h ps://doi.o g/10.1016/j.ij obp.2017.10.022.
[9] Couwenbe g AM, Bu bach JPM, an G e ens ein WMU, e al. E ec o neoadju an
he apy and ec al su ge y on heal h- ela ed quali y o li e in pa ien s wi h ec al
cance du ing he i s 2 yea s a e diagnosis. Clin Colo ec al Cance 2018;17:
e499–512. h ps://doi.o g/10.1016/j.clcc.2018.03.009.
[10] Juul T, Ahlbe g M, Biondo S, e al. Low an e io esec ion synd ome and quali y o
li e: an in e na ional mul icen e s udy. Dis Colon Rec um 2014;57:585–91.
h ps://doi.o g/10.1097/DCR.0000000000000116.
[11] Mo is EJ, Taylo EF, Thomas JD, e al. Thi y-day pos ope a i e mo ali y a e
colo ec al cance su ge y in England. Gu 2011;60:806–13. h ps://doi.o g/
10.1136/gu .2010.232181.
[12] Manceau G, Ka oui M, We ne A, Mo ensen NJ, Hannoun L. Compa a i e
ou comes o ec al cance su ge y be ween elde ly and non-elde ly pa ien s: a
sys ema ic e iew. Lance Oncol 2012;13:e525–36. h ps://doi.o g/10.1016/
S1470-2045(12)70378-9.
[13] Papillon J. In aca i a y i adia ion o ea ly ec al cance o cu e A se ies o 186
cases. Cance 1975;36:696–701. h ps://doi.o g/10.1002/1097-0142(197508)36:
2+<696::aid-cnc 2820360813>3.0.co;2-x.
[14] G´
e a d JP, Ba be N, Gal J, e al. Planned o gan p ese a ion o ea ly T2–3 ec al
adenoca cinoma: a F ench, mul icen e s udy. Eu J Cance 2019;108:1–16.
h ps://doi.o g/10.1016/j.ejca.2018.11.022.
[15] Ge a d JP, Ba be N, Schiappa R, e al. Neoadju an chemo adio he apy wi h
adia ion dose escala ion wi h con ac x- ay b achy he apy boos o ex e nal beam
adio he apy boos o o gan p ese a ion in ea ly cT2–cT3 ec al adenoca cinoma
(OPERA): a phase 3, andomised con olled ial. Lance Gas oen e ol Hepa ol
2023;8:356–67. h ps://doi.o g/10.1016/S2468-1253(22)00392-2.
[16] Papillon J, Be a d P. Endoca i a y i adia ion in he conse a i e ea men o
adenoca cinoma o he low ec um. Wo ld J Su g 1992;16:451–7. h ps://doi.o g/
10.1007/BF02104446.
[17] Schild SE, Ma enson JA, Gunde son LL. Endoca i a y adio he apy o ec al
cance . In J Radia Oncol Biol Phys 1996;34:677–82. h ps://doi.o g/10.1016/
0360-3016(95)02098-5.
[18] Aumock A, Bi nbaum EH, Fleshman JW, e al. T ea men o ec al adenoca cinoma
wi h endoca i a y and ex e nal beam adio he apy: Resul s o 199 pa ien s wi h
localized umo s. In J Radia Oncol Biol Phys 2001;51:363–70. h ps://doi.o g/
10.1016/s0360-3016(01)01677-7.
[19] Dhadda A, Sun Myin A, Thamphya B, e al. A mul i-cen e analysis o adju an
con ac X- ay b achy he apy (CXB) in ec al cance pa ien s ea ed wi h local
excision–P elimina y esul s o he CONTEM1 s udy. Radio he Oncol 2021;162:
195–201. h ps://doi.o g/10.1016/j. adonc.2021.07.021.
[20] Beneze y K, Mon agne L, E esque L, e al. Clinical esponse assessmen a e
con ac X-Ray b achy he apy and chemo adio he apy o o gan p ese a ion in
ec al cance T2–T3 M0: The ime/dose ac o in luence. Clin T ansl Radia Oncol
2020;24:92–8. h ps://doi.o g/10.1016/j.c o.2020.07.001.
[21] Gas oen e ology, F.N.S.o. Diseases and cance s o he diges i e sys em. Rec al
Cance ;2023. <h ps://www.sn ge.o g/ ncd/cance -du- ec um/>[accessed 5
Sep embe 2023].
N.W. Than e al. Radio he apy and Oncology 199 (2024) 110465
8
[22] Aus in PC, Fine JP. The use o p opensi y sco e me hods wi h su i al o ime- o-
e en ou comes: epo ing measu es o e ec simila o hose used in andomized
expe imen s. S a Med 2014;33:1242–58. h ps://doi.o g/10.1002/sim.8008.
[23] Chesnaye NC, S el VS, T ipepi G, e al. An in oduc ion o in e se p obabili y o
ea men weigh ing in obse a ional esea ch. Clin Kidney J 2022;15:14–20.
h ps://doi.o g/10.1093/ckj/s ab158.
[24] Glynne-Jones R, Hughes R. Comple e esponse a e chemo adio he apy in ec al
cance (wa ch-and-wai ): ha e we c acked he code? Clin Oncol (R Coll Radiol)
2016;28:152–60. h ps://doi.o g/10.1016/j.clon.2015.10.011.
N.W. Than e al. Radio he apy and Oncology 199 (2024) 110465
9
