Scandinavian Journal of Surgery 95: 243–248, 2006

Low MaruyaMa Index surgery for gastrIc cancer

s. a. Hundahl
Department of Surgery, U.C. Davis
Surgical Service, VA Northern California Health Care System, Mather, CA, U.S.A.

aBstract

Background: Japanese definitions and treatment guidelines have dominated extent-of-

surgery concepts in gastric cancer for over 4 decades, despite the fact that such defini-
tions/guidelines have changed considerably over time, and the fact they have largely
failed to improve survival in prospective, randomized clinical trials.
Aim: To briefly review lessons from previous surgical trials in gastric cancer, and, more
specifically, to review data validating the concept of “low Maruyama Index surgery” as
a data-driven guide to surgical treatment.
Methods: Review of results from blinded multivariate analyses of two separate, pro-
spective, randomized clinical trials: a) the Macdonald Trial of adjuvant postoperative
chemo-radiation, Intergroup 0116, conducted in North America; and b) the Dutch D1–D2
Trial.
Results: Blinded univariate and multivariate analysis of both trials establish
“Maruyama Index of Unresected Disease” (MI) < 5 as a strong independent predictor of
better disease-free and overall survival in gastric cancer. Moreover, a strong “dose re-
sponse” effect for MI versus survival is apparent.
Conclusions: In contrast to surgery focused on achievement of a particular Japanese-
defined D-level, “low Maruyama Index surgery” is associated with increased disease-free
and overall survival. Further, the dose-response effect suggests MI can be used to quan-
tify the adequacy of lymphadenectomy for a given patient. Low MI surgery can be pro-
spectively planned by using the Maruyama Computer Program pre-operatively or intra-
operatively.
Key words: “Maruyama Index”; “low MI”; gastric; cancer; surgery; lymphadenectomy

INtroDUCtIoN on lawn and garden. A few arrogant weeds on the

green have emerged, testimony to the capacity of na-
As summer sunlight punctuates features of my yard, ture to thwart any human-imposed order. Does the
wafting through leaves of my beloved Japanese ma- scene warrant a herbicidal tsunami, capable of sweep-
ples, I reflect that I’m not immune to seasonal fixation ing each weed-spore to its biological doom? Surgeons
adhering to the Japanese “General rules” du jour for
gastric cancer treatment would have me localize the
primary source, and, based simply on location, spray
Correspondence: like a texan to cover all downwind sod stations and
Scott Hundahl, M.D.
Department of Surgery, U.C. Davis
then some. An alternative approach, taking advan-
Sacramento VA at Mather tage of experience with previous yards, and with pre-
10535 Hospital Way (112) vious weeds, might be driven by consideration of not
Mather, CA 95655-1200, U.S.A. just location, but also the depth of infestation, the
Email: scott.hundahl@va.gov characteristics of dandelions versus crabgrass, and
244
Prospective randomized clinical trials – total vs. subtotal gastrectomy -
total vs. subtotal trials Inclusion N Mortality / Survival
criteria
Gouzi et al.1 Antral tumor
M-0
bozetti et al.2, 3 > 6 cm proximal
margin possible
M-0
robertson et al.4 Antral > 6 cm margin
M-0, Age < 75
other factors. this might permit a more focused, more
effective, solution, addressing the sod at risk while
leaving the rest of the tender turf to pristine contin-
ued health. As we shall see, computerized searching
of comprehensive treatment/outcome records de-
rived from actual treatment of actual gastric cancer
patients, permits the latter, more focused approach
for the treatment of gastric cancer. And someday
soon, I suspect something very similar will be avail-
able for my yard.
rEVIEW of proSpECtIVE rANDoMIzED
trIAlS ADDrESSING roUtINE totAl
VS SUbtotAl GAStrECtoMy
table 1 summarizes prospective, randomized trials of
routine total gastrectomy versus subtotal gastrectomy
for lesions which can be successfully cleared to nega-
tive margins by either procedure (1–4). the robertson
trial, conducted in Hong Kong, includes the addition
of extended node dissection to the total gastrectomy
arm (4). As seen in the table, routine total gastrectomy
fails to generate a survival advantage. In the robert-
son trail, the “subtotal + D1” group actually enjoyed
superior survival compared to the “total + D3 group”
(4).
CHANGING lyMpHADENECtoMy
DEfINItIoNS – tHE DIffICUlt “D”
Gastric cancer lymphadenectomy has been histori-
cally defined according to (several variations of) Jap-
anese definitions. Such definitions reflect various
treatment mandates contained in various editions of
Japanese standardized treatment/staging rules, dat-
ing from 1963 to present (5–8). the Japanese treat-
ment and classification system has, since its incep-
tion, included numeric designations for various
lymph node stations and sub-stations around the
stomach, 31 at last count (8). Definitions for various
nodal levels, originally N1 through N4, are expressed
in terms of groupings of these numbered anatomi-
cally-defined nodal stations for tumors in various
positions within the stomach. the N groupings have
S. A. Hundahl
tAblE 1
Mortality / Survival p Value
(Survival)
SUb-totAl totAl
3% / 48% 1% / 48% n.s.
(5-yr survival) (5-yr survival)
1% / 65% 2% / 62% n.s.
(5-yr survival) (5-yr survival)
Sub-total + D-1 total + D = 3
0% / 1511 dys 3% / 922 dys 0.04
median surv. median surv. 0.07
changed considerably over time. Current definitions
include only node levels from N1 thorough N3 (i.e.
the N4 node grouping, present in editions 1–12, has
now disappeared). With the 13th edition of the Japa-
nese classification system, circa 1997, which has been
translated into English (9), several major changes
were instituted. for example, one major change in-
volves the designation of node involvement at peri-
gastric short gastric (#4a) and left paracardial (#2)
sites as distant metastatic disease sites (M1) when
they occur in the setting of an antral primary tumor.
the same stations are N1 disease or N3 disease for
other primary sites within the stomach. the reader
should appreciate that the Japanese system bears
scant relation to the more familiar UICC/AJCC tNM
system (10).
Japanese mandates for node dissection are classi-
fied according to the “D-level” system. Definitions for
extent-of-lymphadenectomy in all but one of the tri-
als we will discuss follow the Japanese mandates.
According to the Japanese system, a “D1 lymphade-
nectomy” encompasses all anatomically-defined N1
node stations for a given location of tumor, a “D2” all
N2 nodes stations, and a “D3” all N3 node stations.
to make matters a bit more confusing, all but one of
the trials we shall discuss use D-level definitions
based on pre-1997 editions.
proSpECtIVE rANDoMIzED trIAlS – D1
VErSUS D2 or D2+
table 2 summarizes prospective randomized trials of
various Japanese-defined lymphadenectomy schemes
(11–19). In view of the aforementioned changing def-
initions, and resulting confusion, surgeons might
consider themselves fortunate that the trials are all
largely negative.
for the two large European trails, the Cuschieri
trial, conducted by the MrC in britain (13–14) and
the bonenkamp-Hartgrink trial, conducted in the
Netherlands (15–17), in-hospital surgical mortality
for the D2 groups was quite high, 13% and 10%, re-
spectively. Both trials showed significantly higher
mortality when pancreatic-splenic resection was per-
formed, and this somewhat confounded the D-level
 Low Maruyama Index surgery for gastric cancer 245

tAblE 2
Prospective randomized clinical trials. Lymphadenectomy according to Japanese treatment methods (D-level)

lymphadenectomy trials Inclusion N Mortality / Survival Mortality / Survival p value

criteria (Survival)
D-1 D-2
Dent11, 12 t1-3;N0-1;M0 043 0% / 78% 0% / 76% n.s.
Age < 75 (3-yr survival) (3-yr survival)
Cuschieriet al.13,14 Stage I–III 400 6% / 35% 13% / 33% n.s.
Age > 20 (5-yr survival) (5-yr survival
note: non-Japanese definition of “D1” and “D2” for this trial
bonenkampet al.15, 16 Stage I-II 711 4% / 45% 10% / 47% n.s.
Hartgrink et al.17 Age < 85 (5-yr survival) (5-yr survival)
D-1 D-3
Wu et al.18 >t2, M0
Age < 75 156 0% / 45% OS 0% / 51% OS 0.056 – borderline
No esophageal involvement / 49% DfS / 54% DfS 0.15 – n.s.
D-2 D-3
Sasako et al.19 Deep T2–T4 523 0.8% / 69% OS 0.8%/ 70% OS n.s
M0, Age < 76 (5-yr survival) (5-yr survival)

question since, at the time of these trails, these were
mandated procedures for the D-2 group when tumors
were in the middle third or proximal third of the
stomach. In the Dutch trial, restricting subgroup
analysis to patients who did not undergo pancreatic
or splenic resection (a post-hoc, selected analysis),
survival was higher for the D-2 group (59% for the
D-1 group vs. 71% for the D-2 group, p = 0.02) (17).
An 11-year follow-on report for this trial indicates
that of the 12% of cases with pathologic N-2 disease
(N = 89 out of 711 total) , there were nine 10-year sur-
vivors, and eight of the nine were in the D-2 group
(p = 0.01 for this post-hoc analysis of the N-2 sub-
group)(17). Subgroup analysis notwithstanding, both
European trials were negative. Whatever was sup-
posedly gained as a result of D-2 lymphadenectomy
was lost as a result of higher surgical mortality.
In the mid-90’s, as a result of key work by Ma-
ruyama and others (20–22), the Japanese abandoned
routine pancreatic-splenic resection unless required
to achieve a negative-margin resection, and changed
published D2/D3 recommendations accordingly, but
too late for the aforementioned trials. final results
from the above trials verified the wisdom of this
change.
MArUyAMA INDEx of UNrESECtED DISEASE
(MI)
In the late 80’s, Keiichi Maruyama and colleagues at
the National Cancer Center Hospital in tokyo created
a computer program (known as the “Maruyama pro-
gram”) which searched a meticulously-maintained
3,843-patient database of gastric cancer cases treated
by extensive lymphadenectomy (i.e. D2 or more). the
program is designed to match cases with characteris-
tics similar to a given case, and report observed nodal
dissemination risk, survival, and other information.
With seven demographic and clinical inputs, all iden-
tifiable pre-operatively or intra-operatively, the pro-
gram predicts the statistical likelihood of nodal dis-
ease for each of the 16 main nodal stations around the
stomach. Maruyama program predictions have been
assessed in Japanese, German, and Italian popula-
tions and found to be highly accurate (23–25). the
Maruyama program is designed to be used by sur-
geons pre-operatively or intra-operatively, as a con-
venient means of rationally planning a more data-
driven extent of lymphadenectomy for a given pa-
tient. Since the late ‘80’s, the program has been used
in exactly this way by many gastric cancer surgeons
around the world. In an effort to expand use of this
computerized tool, a CD-roM with expanded case
volume was prepared in 2000 (26).
loW-MArUyAMA-INDEx SUrGEry IS
ASSoCIAtED WItH IMproVED SUrVIVAl
IN INtErGroUp 0116, A lArGE U.S. trIAl
of ADJUVANt CHEMo-rADIAtIoN
IN GAStrIC CANCEr
In a prospectively-planned surgical analysis of a large
adjuvant chemo-radiation trial in the U.S. (aka the
“Macdonald trail,” SWoG 9008, Intergroup 0116), the
extent of surgical treatment was specifically assessed
and prospectively coded. the prospectively-planned
surgical analysis of survival made use of a novel
means of quantifying the adequacy of lymphadenec-
tomy relative to likely extent of nodal disease, the
“Maruyama Index of Unresected Disease” (MI) was
defined (by the author, SH) as the sum of Maruyama
Program predictions for those Japanese-defined re-
gional node stations (stations #1–#12) left in situ by
the surgeon (27). Based on the trial’s entry criteria,
and the definition of MI, every case registered could
have had an MI of zero; this variable was under the
surgeon’s control. As depicted in Fig. 1, median over-
246 S. A. Hundahl
MI < 5 62 43 35 28 18
MI ≥ 5 291 263 166 103 55
all survival for the MI < 5 subgroup was 91 months vs
27 months for others (p = 0.005). By multivariate anal-
ysis, adjusting for treatment, t-stage and number of
nodes positive, “Maruyama Index” (MI) proved an
independent predictor of survival (p = 0.0049). Data
for disease-free survival was similar (27–28). Some
impact of “dose of surgery,” as measured by MI, was
also evident: median survival was 20 months for the
highest MI quartile and 46 months for the lowest MI
quartile (treatment-adjusted p = 0.002) (27).
loW-MArUyAMA-INDEx SUrGEry IS
ASSoCIAtED WItH IMproVED SUrVIVAl
IN blINDED rE-ANAlySIS of tHE DUtCH
D1–D2 trIAl
to further assess of the utility of Maruyama Index
(MI) as a prognostic tool, the Dutch D1 vs D2 trial
has recently been re-analyzed (29). blinded to sur-
vival, and eliminating cases with incomplete infor-
mation, 648 of the 711 patients treated with curative
intent had MI assigned. Median MI was 26 (vs. me-
dian of 70 for the Macdonald Trial). Overall trial find-
ings with respect to D-level were not affected by the
absence of the 63 cases with incomplete data (i.e. no
survival difference between D1 and D2). In contrast
to D level, MI < 5 proved a strong predictor of sur-
vival by both univariate and multivariate analysis
(see fig. 2). MI was an independent predictor of both
overall survival (p = 0.016, HR = 1.45, 95% CI 1.07–
1.95) and relapse risk (p = 0.010, HR = 1.72, 95% CI
1.14–2.60). Strong “dose-response” with respect to MI
and survival was also observed (see fig. 2). thus, the
Dutch Trial findings with respect to MI largely con-
firmed what was observed in the Macdonald Trial.
fig. 1. Updated overall survival
curves for Maruyama Index (MI)
< 5 vs ≥ 5 in the Macdonald
chemo-radiation trial (Intergroup
0116). reprinted courtesy of
World Journal of Surgery (ref 29).
With kind permission of Springer
Science and business Media.
4 0
13 0
DISCUSSIoN
based on results from the aforementioned trials, it
appears that surgeons might better impact on patient
survival by pursuing a “low Maruyama Index opera-
tion” instead of relying on D-level guidance. by using
the Maruyama program to prospectively plan a given
patient’s lymphadenectomy, achieving a “low
Maruyama Index operation” is relatively straightfor-
ward. And in the current era, what surgeon or operat-
ing room does not have access to a laptop or pC to
display the output?
the compelling dose-response effect for MI also
suggests it can also be viewed as a quantitative yard-
stick for the adequacy of lymphadenectomy in a
given case of gastric cancer. As such a quantitative
yardstick, it might someday be used to perhaps iden-
tify patients at greater or lesser risk of local-regional
recurrence, and perhaps influence decisions on post-
operative adjuvant therapy. At a minimum, MI should
be explicitly calculated and reported for every patient
entered into a postoperative adjuvant trail.
SUMMAry
With negative-margin resection – achieved by distal
subtotal gastrectomy when feasible, and with pan-
creas and spleen preserved unless invaded – lymph-
adenectomy can now be customized by pre-operative
or intra-operative use of the Maruyama program to
generate a “low-Maruyama-Index operation,” thus
optimizing survival by restricting dissection to node
stations at risk for disease. the compelling dose-re-
sponse effect for MI also suggests it can also be
viewed as a quantitative yardstick for the adequacy
 Low Maruyama Index surgery for gastric cancer
fig. 2. overall survival for various MI quartiles
in the Dutch D1-D2 trail. reprinted courtesy
of World Journal of Surgery (ref 29). With kind
permission of Springer Science and business
Media.
of lymphadenectomy in a given case of gastric cancer.
future adjuvant trials in gastric cancer should report
this variable for every patient registered.
rEfErENCES
01. Gouzi Jl, Huguier M, fagniez pl, launois b, flamant y,
lacaine f, paquet JC, Hay JM: total versus subtotal gastrec-
tomy for adenocarcinoma of the gastric antrum. A french pro-
spective controlled study. Ann Surg 1989;209(2):162–166
02. bozzetti f, Marubini E, bonfanti G, Miceli r, piano C, Crose N,
Gennari l: total versus subtotal gastrectomy: surgical morbid-
ity and mortality rates in a multicenter Italian randomized
trial. the Italian Gastrointestinal tumor Study Group. Ann
Surg 1997;226(5):613–6120
03. bozzetti f, Marubini E, bonfanti G, Miceli r, piano C, Gennari
L: Subtotal versus total gastrectomy for gastric cancer: five-
year survival rates in a multicenter randomized Italian trial.
Italian Gastrointestinal tumor Study Group. Ann Surg 1999;
230(2):170–178
04. Robertson CS, Chung SC, Woods SD, Griffin SM, Raimes SA,
lau Jt, li AK: A prospective randomized trial comparing r1
subtotal gastrectomy with r3 total gastrectomy for antral can-
cer. Ann Surg 1994;220(2):176–182
05. Sayegh ME, Sano t, Dexter S, Katai H, fukagawa t, Sasako M:
tNM and Japanese staging systems for gastric cancer: how do
they coexist? Gastric Cancer 2004;7(3):140–148
06. Kajitani t: the general rules for the gastric cancer study in
surgery and pathology. Part I. Clinical classification. Jpn J Surg
1981;11(2):127–139
07. Japanese Classification of Gastric Carcinoma, First English
Edition. tokyo: Kanehara & Co ltd, 1995
08. Degiuli M, Sasako M, ponti A, Soldati t, Danese f, Calvo f:
Morbidity and mortality after D2 gastrectomy for gastric can-
247
cer: results of the Italian Gastric Cancer Study Group prospec-
tive multicenter surgical study. J Clin oncol 1998;16(4):1490–
1493
09. Japanese Gastric Cancer A. Japanese Classification of Gastric
Carcinoma – 2nd English Edition. Gastric Cancer 1998;1(1):10–
24
10. Greene fl, page Dl, fleming ID, fritz AG, balch CM, Haller
DG, Morrow M (Eds): AJCC Cancer Staging Manual. New
york: Springer; 2002
11. Dent DM: radical surgery for curable gastric carcinoma. S Afr
Med J 1994;84(2):56–57
12. Dent DM, Madden MV, price SK: randomized comparison of
r1 and r2 gastrectomy for gastric carcinoma. br J Surg 1988;
75(2):110–112
13. Cuschieri A, fayers p, fielding J, Craven J, bancewicz J, Joy-
paul V, Cook p: postoperative morbidity and mortality after
D1 and D2 resections for gastric cancer: preliminary results of
the MrC randomised controlled surgical trial. Surgical Co-
operative Group. Lancet 1996;347(9007):995–999
14. Cuschieri A, Weeden S, fielding J, bancewicz J, Craven J, Joy-
paul V, Sydes M, fayers p: patient survival after D1 and D2
resections for gastric cancer: longterm results of the MrC ran-
domized surgical trial. Surgical Cooperative Group. br J Can-
cer 1999;79(9–10):1522–1530
15. bonenkamp JJ, Songun I, Hermans J, Sasako M, Welvaart K,
plukker Jt, van Elk p, obertop H, Gouma DJ, taat CW, et al:
randomised comparison of morbidity after D1 and D2 dissec-
tion for gastric cancer in 996 Dutch patients. lancet 1995;
345(8952):745–748
16. bonenkamp JJ, Hermans J, Sasako M, van de Velde CJ, Wel-
vaart K, Songun I, Meyers S, plukker Jt, Van Elk p, obertop H,
Gouma DJ, van landschot JJ, taat CW, de Graaf pW, von Mey-
enfeldt Mf, tilanus H; Dutch Gastric Cancer Group: Extended
lymph-node dissection for gastric cancer. Dutch Gastric Cancer
Group. N Engl J Med 1999;340(12):908–914
17. Hartgrink HH, Van De Velde CJ, Putter H, Bonenkamp JJ,
Klein Kranenbarg E, Songun I, Welvaart K, van Krieken JH,
248 S. A. Hundahl
Meijer S, plukker Jt, van Elk pJ, obertop H, Gouma DJ, van
landschot JJ, taal CW, de Graaf pW, von Meyenfeldt Mf, ti-
lanus H, Sasako M: Extended lymph node dissection for gas-
tric cancer: Who may benefit? Final
final results of the randomi�ed
randomized
Dutch gastric cancer group trial. J Clin oncol 2004;22:2069–
2077
18. Wu CW, Hsiung CA, lo SS, Hsieh MC, Chen JH, li At, lui
Wy, Whang-peng J: Nodal dissection for patients with gastric
cancer: a randomised controlled trial. Lancet Oncol 2006;7(4):
309–315
19. Sasako M, Sano S, yamamoto A, Nashimoto A, Kurita H: ran-
domized phase III trial of standard D2 versus D2 + para-aortic
lymph node (pAN) dissection (D) for clinically M0 advanced
gastric cancer: JCoG9501. J Clin oncol 2006;24 (18-s, part II):
934-s
20. Maruyama K, Sasako M, Kinoshita t, Sano t, Katai H, okajima
K: pancreas-preserving total gastrectomy for proximal gastric
cancer. World J Surg 1995;19(4):532–536
21. Kasakura y, fujii M, Mochizuki f, Kochi M, Kaiga t: Is there a
benefit of pancreaticosplenectomy with gastrectomy for ad-
vanced gastric cancer? Am J Surg 2000;179(3):237–242
22. Uyama I, ogiwara H, takahara t, Kikuchi K, Iida S, Kubota t,
Kumai K, Kitajima M: Spleen- and pancreas-preserving total
gastrectomy with superextended lymphadenectomy including
dissection of the para-aortic lymph nodes for gastric cancer.
J Surg Oncol 1996;63(4):268–270
23. Kampschoer GH, Maruyama K, van de Velde CJ, Sasako M,
Kinoshita t, okabayashi K: Computer analysis in making pre-
operative decisions: a rational approach to lymph node dissec-
tion in gastric cancer patients. Br J Surg 1989;76(9):905–908
24. bollschweiler E, boettcher K, Hoelscher AH, Sasako M,
Kinoshita t, Maruyama K, Siewert Jr: preoperative assessment
of lymph node metastases in patients with gastric cancer:
evaluation of the Maruyama computer program. br J Surg
1992;79(2):156–160
25. Guadagni S, de Manzoni G, Catarci M, Valenti M, Amicucci G,
et al: Evaluation of the Maruyama computer program accu-
racy for preoperative estimation of lymph node metastases
from gastric cancer. World J Surg 2000;24(12):1550–1558.
26. Siewert Jr, Kelsen D, Maruyama K, feussner H, omote K,
Etter M, Hoos A: Gastric cancer diagnosis and treatment – an
interactive training program. ISBN 354014787X In. 1 ed:
Spinger Electronic Media; 2000
27. Hundahl SA, Macdonald JS, Benedetti J, Fit�simmons T: Surgi-
cal treatment variation in a prospective, randomized trial of
chemoradiotherapy in gastric cancer: the effect of undertreat-
ment. Ann Surg Oncol 2002;9(3):278–286
28. Hundahl SA, Macdonald JS, benedetti J: Durable survival im-
pact of “low Maruyama Index Surgery” in a trial of adjuvant
chemoradiation for gastric cancer. San francisco: 2004 ASCo
GI Symposium; 2004 January 22, 2004. report No.: 48
29. peeters KCMJ, Hundahl SA, Kranenbarg EK, Hartgrink H, van
de Velde CJH: “low-Maruyama-Index” Surgery for gastric
cancer – A blinded re-analysis of the Dutch D1–D2 trial. World
J Surg 2005;29:1576–1584
received: August 28, 2006