Br. J. Surg. Vol.

69 (1982) 613-616 Printed in Great Britain

The mesorectum in rectal cancer

surgery-the clue t o pelvic
recurrence?
Five cases are described where minute foci of adenocarcinoma have
been demonstrated in the mesorectum several centimetres distal to the
apparent lower edge of a rectal cancer. In 2 of these there was no
other evidence of lymphatic spread of the tumour. In orthodox
anterior resection much of this tissue remains in the pelvis, and it is
suggested that these foci might lead to suture-line or pelvic
recurrence. Total excision of the mesorectum has, therefore, been
J. HEALD, M. HUSBAND carried out as a part of over 100 consecutive anterior resections.
AND R. D. H. RYALL Fifty of these, which were classiJied as ‘curative’ or ‘conceivably
Bdsiiigstohe Bowel Cancer Clinic. Basingstoke
curative’ operations, have now been followed for over 2 years with no
District HOFpitdl. Bdsingstoke, Hdmpshire pelvic or staple-line recurrence.
T h e incidence of locally recurrent disease is the most important Line of excision includes mesorectum
measure of the success of any new operation for rectal cancer.
Thus there has been anxiety (1 ) that the increase in sphincter-
conserving surgery due t o staplers might lead t o more local
recurrences. Four years ago, therefore, we combined the
decrease in permanent colostomies in our unit with a change in
the technique for pelvic dissection. In particular we determined
that all cancers of the midrectum should be excised with the
mesorectum intact. Thus the phase of dividing this during
anterior resection, which is described in standard textbooks (2),
was completely omitted and the whole mesorectum was encom-
passed by the plane of excision. In this way none of the usual
‘block‘ of fatty lymphovascular tissue remains in the posterior
half of the pelvis even though the anus, the levators, a small
rectal reservoir and as much as possible of the nerve plexuses
have been preserved.

Operative and histological methods

A full length abdominal incision was made from the xiphisternum to /
Site of tumour deposits in Case 6
the pubis. The plane surrounding the left half of the colon was
developed extensively with careful preservation of the autonomic nerve Fig. 1. The suggested plane of excision is shown diagrammatically by
plexuses. Under direct vision with sharp scissor dissection this plane the dashed line. The remnant from Case 6 which contained an isolated
was extended down into the pelvis around the rectum, the tumour and microscopic deposit is indicated by the arrow.
particularly around the fatty mesorectum as far as the point of
emergence of the anorectum from the levator gutter. Great efforts were
made to avoid digital extraction of the tumour on the grounds that this Frozen section during surgery .for checking mural rleurance
could tear veins or split into tumour planes. The pelvic fascia and the In poorly differentiated tumours or those with less than 2cm distal
autonomic nerve plexuses were preserved by pursuing the lipoma-like clearance the main specimen was sent for frozen section examination of
outer surface of the mesorectum except in the immediate vicinity of the the distal margin of the muscle tube. In some cases a thin slice of rectal
tumour. In most cases the correct plane was avascular except where it margin would fit on one chuck and in others 2 blocks were required.
was crossed by the lateral rectal vessels which could simply be cut and The ‘doughnut’ was examined by paraffin wax section later and was
packed with a small gauze swab. Only occasionally did they require regarded as an absolute minimum extra safety margin over and above a
ligation after the tumour was removed. The inferior mesenteric artery report of a ‘clear edge’ on the main specimen.
was ligated I cm from the aorta and the vein separately 1 cm from the
splenic vein. The proximal lymphatic clearance was thus ‘radical’ Antibacterials
though not to the extreme of a ‘pre-aortic strip’. The anorectum was Every patient in our own unit received pre- and peroperative cepha-
then cross-clamped beyond the tumour if this was possible, or the anus zolin or cephamandole and metronidazole.
was excised from below if it was not. I n the former group the clamp was
drawn up into the wound for incision and the purse-string was inserted Wash-outs during surger.v
anteriorly before the posterior wall was cut. The surgeon had to These were performed with water in every case immediately after the
balance the desirability of mural clearance beyond the tumour against rectum was cross-clamped.
the value of leaving a small rectal reservoir which would undoubtedly Histopathological methods
contribute to subsequent function. We usually judged it safe to leave The specimens were opened and placed in formalin and the sizes of the
2-4cm of rectal stump above the levators so that most of the tumour and the distal margin were measured after fixation-they were
anastomoses lay at 5-8 cm from the anal verge. The distal mesorectum not stretched out. For histological examination the following routine
was dissected off the back of the muscle tube with scissors whilst it was was followed.
drawn upwards with a dry swab-the angle was thus emptied of the last One block was taken at the site of maximal extension through the
remnants of fatty lymphovascular tissue (Fig. 1). The technique for the wall. One block was taken from the distal edge of the specimen and 3-4
actual stapling has been described elsewhere (3). This technique was not tangential blocks from the mesenteric edge for assessment of vascular
necessary for some upper rectal tumours which mobilized up so well invasion (4, 5). A longitudinal block was taken through the lower edge
that they could safely be resected like sigrnoid lesions, i.e. the meso- of the tumour and the adjacent (apparently uninvolved) wall. As many
rectum was transected at least 5 cni below the tumour and the lateral lymph nodes as possible were dissected out of the specimen and
ligaments were not divided. examined.
614 R. J. Heald et al.

Fig. 2. A 3mm node from cut surface of mesorectum afer orthodox Fig. 3. Second anterior resection specimen from a patient with a
anterior resection (Case I I . recurrence (Case 21,

Case 2: A man of 7 1 years who refused abdominoperineal resection was

Fig. 4. Intact mucosa over the tumour deposit (Case 2). referred with a suture-line recurrence 6 months after conventional
Recently 3 blocks from the apparently normal distal mesorectum anterior resection for a moderately differentiated Dukes’ C tumour. He
were examined histologically in relevant cases and in 4 patients tumour underwent a second (low stapled) anterior resection and died 19 months
was identified in the lower mesorectum (Cases 3-6). In Case 6 the later of liver metastases with no evidence of pelvic disease. Fig. 3
deposit was only identified after examination of multiple sections cut at illustrates the volume of mesorectal tissue which remained after the first
levels through the block. Multiple levels are not examined routinely in operation but was encompassed by the second, and histology (Fig. 4)
negative cases but on this occasion there was a lymphocytic infiltrate shows that the bulk of the tumour lay outside the bowel wall with the ‘tip
around and within the walls of blood vessels and lymphatics which may of the iceberg’ in the submucosa under an intact mucosa. This is the type
be a useful pointer to the likely presence of extramural venous or of suture-line recurrence, arising in perhaps lymphatic vessel or node,
lymphatic infiltration in the adjacent tissue. This method is now being which total mesorectal excision aims to prevent.
used more frequently to examine the apparently normal mesorectal
tissue. Cases 3, 4 and 5: The 3 blocks taken from the distal mesorectal tissues
Case reports are now serial sectioned and in 3 further anterior resection cases tumour
Case 1: A woman of 49 years underwent conventional anterior resection has been demonstrated in small lymphatics. In 1 of these cases tumour
for a carcinoma at 8 cm in 1973.Careful histological examination of the was present in the first sections examined and would therefore have been
bowel showed no distal spread of tumour, but a single unremarkable detected by routine sectioning. In the other 2 cases tumour was only
lymph node (3mm in diameter) on the surface of the divided detected in sections cut from deeper levels and these would, by routine
mesorectum approximately 3 cm below the tumour edge revealed a tiny sectioning, have been reported as negative.
focus of tumour within it (Fig. 2). Two cases were poorly differentiated. One of these showed several
If it had not been for this discovery the tumour would have been lymph node metastases close to the base of the tumour and the
classified as a moderately differentiated Dukes’ B lesion with slight extramural spread in this case was moderate. The other case also
extrarectal spread and vascular invasion. After discussion between the showed several lymph nodes involved close to the tumour but the
authors, the case was converted to an abdominoperineal excision extramural spread was moderate. In Case 3 the tumour, a moderately
though no further tumour was found. She progressed well for 6 years well-differentiated lesion, had slight extramural spread and only 1
until a terminal phase when bony secondaries led to her death. No lymph node in the tumour base contained a metastatic deposit. In this
recurrence of pelvic disease was ever detected clinically but we case the deposit in the distal mesorectum was over 4cm from the
postulate that the lymphatic micrometastasis could well have led to extramural tumour. In the other 2 cases the deposit in the distal
local recurrent tumour if it remained. Whether or not this is true, the mesorectum was 1-2cm from the extramural tumour.
case illustrates the potential of rectal cancer for distal lymphatic spread In none of the 3 cases was a tumour found in the lower gun ring or in
in the absence of intramural spread. the distal rectal wall.
Clue to pelvic recurrence 615

Table I: 65 CONSECUTIVE ANTERIOR RESECTIONS

5 deaths (all over 70yr) I
60 patients for follow-up
I
10 after ‘palliative’ surgery I
50 after ‘curative’ or ‘conceivably curative’ patients 1
I
Table 11: I0 PALLIATIVE OPERATIONS
8 had liver secondaries
I tumour ruptured during dissection
1 had microscopic tumour in the doughnut

Table III: MARGIN (UNSTRETCHED)

Margin
(em) ‘Curative’ ( n = 50) Polliutiw In = 10)

4-3 9 2.
3-2 II 1
2- 1 10 3
0- 1 10 2

Fig. 6. Suhmucosal deposit in the ‘doughnut’correctly predicted the

Table IV: RECURRENCE OF CANCER IN SO only staple-line recurrence.
PA TIENTS
(unstretched). The average tumour height above the anal verge was
Liver secondaries (n = 4; 8%) 3 dead (6%) 8.9cm and the anastomotic height 5.4cm. Thus the tumours are
1 alive and well 2 + yr after somewhat lower than the previously published anterior resection series,
hepatic trisegmentectomy including 29 with a sigmoidoscopic height of 8 cm or less and 35 which
Lung/liver/bone n = 2 (4””) we consider would formerly have received abdominoperineal excisions.
Pelvic wall None
Suture line None
Results
Figures are for minimum 2 yr/mean 3 yr postoperation. Four of the patients have died of unconnected causes. Only 3
have died of cancer, and 2 others are alive with metastatic
Case 6; A 79-year-old man underwent a low stapled anterior resection. disease (Table IV). Thus only 10 per cent of the patients have so
A narrow pelvis made total mesorectal excision difficult and a 2 x 1 cm far developed metastases and careful examination of the pelvic
distal remnant was missed in the main dissection but subsequently wall has revealed no clinical evidence of disease on the suture-
excised from the ‘cul-de-sac’ between the anorectal stump and the line or the pelvic wall. Nine out of 10 of the palliative cases have
levator gutter (Fig. I). The main specimen showed a 5cm Dukes’ B died and 3 of these had evidence of pelvic disease at their death.
tumour with only 0.5cm extramural extension and neither venous nor ,One died of uraemia after an operation leaving obvious tumour
lymph node involvement. There was 4 cm of clear bowel margin and 10 on the pelvic wall. A second died of generalized disease after an
nodes had been examined. Later extensive examination of multiple operation where the tumour had ruptured. A third died of
levels through the separate remnant revealed a tumour deposit 0.5 mm
in diameter surrounded by lymphocytes, probably lying within a generalized disease also manifesting a staple-line recurrence.
lymphatic vessel (Fig. 5). This case is particulary important because this This had been predicted because a focus of tumour had been
solitary microscopic satellite was 4cm distal to the apparent lower edge observed in a submucosal lymphatic in the ‘doughnut’ (Fig. 6).
of the tumour. No further surgery had been attempted because 17 out of 19
nodes were involved and because the patient refused
Follow-up series colostomy.
Between April 1978 and January 1982 113 consecutive anterior
resections for cancers between 4 and 15cm from the anal verge have
been carried out with total mesorectal excision as described. Discussion
Abdominoperineal excision has been reserved for tumours extending There are good reasons for suspecting that distal spread of
too close to the anus for mural clearance or safe cross-clamping at rectal cancer is often initially confined, as in these cases, to the
surgery: it was rarely performed because of size and fixity within the mesorectal tissues. First, local suture-line recurrences are most
pelvis. Thus the operation has become our standard procedure for most commonly found posteriorly and only from tumours of the
rectal cancers, accounting for more than 4 out of 5 of the operations rectum rather than the colon (6) with the highest incidence in
performed during this period. Only 2 cases have been rejected for the lowest tumours (9). It is only in these that lymphatic tissue
excision of the primary (operability rate 98.2 per cent). is likely to be left behind in the region of anastomosis, which
We consider that a 2-year follow-up has some validity in the might also explain why the pull-through operation seemed
assessment of locally recurrent disease. Fourteen out of E.S.R. Hughes’
19 cases reported by Hardy (6) recurred within I year, and Goligher et immune to suture-line recurrence in Hardy’s series (6) where
al. (7), Tyndal et al. (8) and Morson et al. (9) all emphasize that 80-90 the mesorectum is likely to be excised to make the pullthrough
per cent of local recurrences present within 2 years. In this series, in possible. Secondly, the initial tumour is almost invariably a
January 1982, 65 operations had been performed more than 2 years Dukes’ C lesion with evident lymphatic deposits (Case 2) (9).
previously. Thus 60 operation survivors are available for follow-up None of these observations can readily be explained if
(Table I)with a minimum of 2 years, a mean of 3 years and a maximum implantation or direct intramural spread were the basis for the
of 46 months. Ten of these were classified as ‘palliative’ procedures recurrence. Recent work by Rosenburg (10) suggest that
(Table I I ) and the remaining 50 were either ‘curative’ or ‘conceivably desquamated cells are seldom a viable source of implantation
curative’-the latter description usually reflecting unusual size and
fixity within the pelvis. which is probably rather an uncommon event, although the
Of the 50 curative or conceivably curative patients, 8 were in Dukes’ wash-out is properly established on bacteriological grounds
stage A, 32 in stage B and 10 in stage C; 9 had good differentiation, 31 and may also guard against implantation. Significant spread
had average and 10 had poor. Table III shows that 31 out of the 50 along the muscle tube is also uncommon, except in poorly
had margins along the muscle tube with a length of 3cm or less differentiated lesions (1 1) and it is interesting that rectal stump
616 R . J. Heald e t al.

recurrence was seldom seen after Hartmann’s operation (12). to surgery. This important aspect of staging forms the basis of a
Our experience confirms that the 5 cm margin may probably be further study but the relevant 50 cases are essentially consecu-
safely reduced to 2 cm in differentiated lesions and monitored tive and include many poorly differentiated and fixed lesions
by frozen section in doubtful cases. The only suture-line which would formerly have been excluded.
recurrence in this series was in a palliative case and was Since no adjuvant chemo- or radiotherapy was used, these
correctly predicted by histological examination of the ’dough- promising early figures suggest at the very least that local
nut’ (Fig. 6). Intramural spread is in our opinion, therefore, less recurrence is directly related to the technique of pelvic dis-
dangerous and less common than distal microscopic lymphatic section. It is precisely from consideration of surgical technique
spread within the mesorectum (such as Cases I , 3-6) which that the idea of total mesorectal excision developed-its division
could not possibly have been detected before or during surgery, is bloody but the plane around it is avascular and it is surgically
or even (as in Cases 1 and 6) suspected after routine histological possible to encompass it without undue difficulty. It leads to a
examination. clean muscle tube which is suitable for a purse-string, and if
These findings seem somewhat at variance with Goligher and care is taken, the pelvic fascia and autonomic nerves can often
Dukes’ (7) low figure of 6.9 per cent for distal lymphatic spread be preserved outside it and a small anorectal remnant retained
with only 2 per cent beyond 2cm. Node sampling, however, to provide reasonable function. We suggest that complete
would miss tiny vessel deposits such as in Case 6, and Gilchrist removal of the mesorectum encompasses the most dangerous
(13) demonstrated that the incidence of node involvement rises and prevalent field of spread and its excision is as logical as that
to an astonishing two-thirds in ‘curative’ and 100 per cent in of any mesentery in close proximity to a cancer.
‘palliative’ cases if sufficient time is spent looking for it. Thus
there is plentiful circumstantial evidence that the doctrine of
preferential upward spread should no longer be accepted as the Acknowledgement
entirely safe dictum that it has in the past. This work was supported by a grant from Whittaker Life Sciences.
This complete ‘emptying’ of the pelvis has not, however,
been without problems: we have, despite considerable stapling
experience, been unable to eliminate anastomotic leakage
completely (14). For anastomoses below 6 cm our clinical References
leakage rate has been 13 per cent and this figure rises to 18 per I. Leading Article: Colorectal surgery-the Cinderella specialty. Br.
cent if X-ray detected sinuses are included. It is a reasonable Med. J . 198I ; 283.
2. Rob C. and Smith R.:Operative surgery Vol. 4 Colon, Rectum
hypothesis that the large cavity created by total mesorectal and Anus. 3rd ed. London: Butterworths. 147.
excision may contribute to this problem by encouraging peri- 3. Heald, R. J.: A new approach to rectal cancer. Br. J . Hosp. Med.
anastomotic or posterior collections which point into the lumen 1979: 217.
and cause a ‘late leak’. In addition we preserve where possible 4. Talbot I. C.. Ritchie S., Leighton M. H. et al.: The clinical
2-4cm of rectum which is unsupported by vascular tissue significance of invasion of veins by rectal cancer. Br. J . Surg. 1980;
67: 439-42.
which has been peeled out with the specimen: its blood supply 5. Blenkinsopp W. K., Stewart-Brown S.. Blesovsky L. et al.:
might thus be impaired, although this has never been apparent Histopathology reporting in large bowel cancer. J . Clin. Path
to the surgeon. For these reasons it is now our routine to 1981: 34: 509-13.
protect all anastomoses below 7cm with a loop transverse 6. Hardy K. J . . Cuthbertson A. M. and Hughes E. S. R. Suture-line
colostomy for 5-6 weeks. neoplastic recurrence following large bowel resection. Aust. N Z J .
In assessing our results it is difficult to establish an acceptable Surg. 1971; 41: 44.
7. Goligher J. C., Dukes C. E. and Bussey H . J. R.: Local recurrence
‘norm’ for local recurrence. Anderson (15) reviewed the litera- after sphincter-saving excisions for carcinoma of the rectum and
ture in 1981 and gave the 5-year incidence after restorative rectosigmoid. Br. J . Surg. 1951; 39: 199-211.
excision as ‘12-20 per cent in reports from special interest 8. Tyndal E. C., Docherty M. B. and Waugh J. M.: Pelvic recurrence
centres and about 40 per cent in general surgical practice’. of Ca. rectum. Surg. Gynecol. Obstet. 1964; 118: 47-51.
Since all authorities (6-9) agree that 80-90 per cent are 9. Morson B. C.. Vaughan ,E. G. and Bussey H. J. R.: Pelvic
recurrence after excision of the rectum for carcinoma. Br. Med. J .
apparent at 2 years this preliminary report on 50 cases with a 1963: 2 13-18.
zero incidence after 2-4 years seems already significant and 10. Rosenberg I. L.: The aetiology of colonic suture-line recurrence.
raises the possibility that favourable selection factors might Ann. R . Coil. Surg. Engl. 1979; 61: 251-7.
have contributed. The series has an apparently low incidence of I I. Williams N. S., Dixon M. and Johnston D.: Distal intramural
Dukes’ C lesions relative to the Bs with the usual level of As. spread and local recurrence of rectal cancer. Br. J . Surg. 1979; 66:
890.
The selection policy has been to exclude no cases where 12. Dixon C. F.: Carcinoma of the recto-sigmoid: resection without
excision was physically possible: even tumours occupying the permanent colostomy. Proc. Staff Meet. Ma.vo Clinic 1936; 11:
whole pelvis and fixed within it have been managed by anterior 384.
resection if the anorectum was not invaded. The proportion of 13. Gilchrist R. K.: Lymphatic spread of carcinoma of the colon. Dis.
C cases has now risen in more recent cases to over 50 per cent Colon Rectum 1959: 2: 69-77.
with a continuing low incidence of local recurrence (2/50). 14. Heald R. J. and Leicester R. J.: The low stapled anastomosis. Br.
J . Surg. 1981: 68: 333-7.
We consider therefore that the change in the ratio of Bs to Cs 15. Anderson J. M.: Chemotherapeutic prevention of local recurrence
reflects greater laboratory time and specialized histopatho- after stapled anastomoses in rectal cancer. Scot. Med. J . 1981; 26:
logical interest rather than a true alteration in the cases coming 21-3.