Oncologists accept that the prognosis of a patient with cancer is primarily dependent upon two factors.  First, the aggressive nature of the malignancy is of paramount importance.  If the cancer is of an invasive type, the prognosis is diminished.  If it is non-invasive but rather expands by pushing into the surrounding tissues, the prognosis may be very good even though the cancer is large.

Also, the stage of the malignancy at the time of diagnosis is of paramount importance.  If the cancer has progressed so that the lymph nodes are involved, the prognosis is reduced.  If distant metastases has developed before treatments are initiated, the situation may be hopeless.  

Granted, the aggressive nature and the stage of the cancer are important.  However, for gastrointestinal malignancy, there may be an even more crucial factor in regards to prognosis.  
This is the surgical skill of the responsible surgeon.  
The grim reality is that the difference in survivorship obtained by the best and the worst gastrointestinal surgeons is at least 50%!  
This concept of cancer prognosis has not been adequately explored in recent years.  

DATA FROM THE SURGICAL
LITERATURE

The concept that surgical skill is a crucial prognostic factor is not new.  In 1967, Turnbull and colleagues from the Cleveland Clinic in Ohio, published a manuscript presented at the American Surgical Society on “No-touch isolation techniques.”  They were convinced that the surgeon's gentle handling of tissues would prevent dissemination of cancer cells through the portal system to the liver.(1)  
Turnbull's statistical evaluation of his data may not be acceptable by present day standards; also his hypothesis regarding the mechanism of cancer dissemination is not compatible with current data regarding the natural history of gastrointestinal cancer.  
However, it remains clear that his results with the surgical treatment of colorectal cancer obtained by no-touch techniques were far superior to other reports in the 1960s.  
From England, Phillips and colleagues called attention to the marked variation in the incidence of local recurrence of rectal cancer when the statistics for individual surgeons were tabulated.  
This paper published in the British Journal of Surgery clearly established their opinion regarding surgical skill and its effect on local cancer recurrence.(2)  Ultimately, there would also be a marked difference in survival with or without locally recurrent rectal cancer.  

I described numerous technical factors which will influence survival in colon and rectal cancer surgery in a monograph in 1984.(3)  
The components of an adequate colon or rectal cancer resection were explicitly laid out.  
Again, in 1995, our opinions regarding surgical technique and its impact on local recurrence and survival were summarized.(4)
Hermanek and colleagues in a landmark publication courageously documented the great differences in local recurrence rates when the data was presented on a surgeon by surgeon basis.  
These authors documented the incidence of local recurrence as high as 55% and as low as 5%.(5)  As expected, the survival was directly correlated with the rate of local recurrence. One surgeon had a survival rate as low as 35% and other surgeons had survival rates as high as 85%.  This manuscript looked for risk factors that would account for the marked differences in local recurrence and survival.  The stage of the malignancy and clinical factors involved in its presentation did not seem to explain these great differences. Rather, the surgical skill possessed by the responsible surgeon was thought to result in the great difference in survival.  

Recently, Porter and colleagues from Alberta, Canada reported in the Annals of Surgery the differences in local recurrence and in survival for rectal cancer between two groups of surgeons.(6)  
One group had advanced training and greater experience in the operating room. The other group had no special training and performed their rectal cancer surgery on an occasional basis.  In the experienced and high volume group, the overall survival was 67%.  In the occasional surgeon without additional training, the survival was only 35%.

DISSEMINATION 
OF GASTROINTESTINAL
CANCER 

We know that gastrointestinal cancer disseminates in three ways. It can metastasize via the portal bloodstream to the liver. It may be difficult or impossible for the surgeon to reduce the incidence of metastases within the liver.  Unless the surgeon is extremely rough it is unlikely that he will squeeze cancer cells into the portal blood.  However, theoretically this is possible.  However, gastrointestinal cancer also spreads to lymph nodes and may disseminate to peritoneal surfaces.  If patients recur with lymph node involvement or with progressive disease at the resection site or on peritoneal surfaces, this is the surgeon's responsibility.  Recurrent disease at either of these two sites indicates insufficient skill of the responsible surgeon.  

The hideous flaw that exists with inadequate gastrointestinal cancer surgery can be formulated as follows.  Many patients come to the surgeon with a contained malignancy.  It may be advanced and beginning to disseminate local-regionally; however, as yet there is no disseminated disease.  Unfortunately, in a large proportion of gastrointestinal cancer operations, with unskillful resection the patient leaves the operating room with persistent cancer in lymph nodes or with a disseminated malignancy on the internal lining of the abdomen for which there can be no cure.  This hideous flaw, inadequate surgical skill, is not currently anticipated by those who must undergo gastrointestinal cancer resection.
 Unfortunately it is very possible that they will die, not because their cancer was aggressive or was at a late stage; rather they may die because of the responsible surgeon's lack of knowledge and/or skill.

CONTAINMENT THROUGH
CENTRIPETAL SURGERY

What are the essential components of adequate gastrointestinal cancer surgery? These can be listed as follows: wide exposure of the operative field, absolute hemostasis, adequate lateral margins of dissection, adequate lymph node dissection, and knowledgeable use of perioperative intraperitoneal chemotherapy.  If the surgeon is to perform surgery optimally, he must clearly visualize the primary cancer.  There should never be a struggle for adequate visualization of the cancer or of the surrounding abdominal or pelvic viscera.  A large midline abdominal incision extending from xiphoid to pubis is required.  Self-retaining retractors are employed to hold back the edges of the abdominal incision.  Frequent irrigation is performed in order to clear the operative field of blood.  A second tier of retractors is often necessary to prevent repeated manipulation and repositioning of the bowel in order to provide exposure. This is especially true when working deep in the pelvis (rectal cancer) or high up under the diaphragms (gastric cancer).  Finally, there are some more difficult surgical procedures where the help of a resident is not adequate.  Some difficult cases demand two experienced surgeons who have worked together frequently in the past.  When the case is very difficult the most skillful assistant available is absolutely necessary.  

If adequate hemostasis is to be achieved, the surgeon must give up scissor and knife dissection.
Dissection should be performed with lasermode electrosurgery in order to keep blood loss to a minimum.(7)  Continued ooze from small bleeding points should not be tolerated during the dissection.  Of course, all large blood vessels are ligated in continuity before they are divided.  Morsillation of fat surrounding the medium and large blood vessels by pressure of the thumb and index finger will facilitate clean ligation in continuity of blood vessels.

The procedural dictum for adequate containment of the gastrointestinal surgery is called CENTRIPETAL SURGERY.  
In this approach to cancer resection one must move around the tumor mass with perfect hemostasis, adequate margins of dissection, and sufficient visualization so that vital structures are not damaged.  If all of these requirements are not met, the surgeon must attack the malignant disease from another anatomic site.

Centripetal surgery insures adequate margins of normal tissue left intact and without trauma around the cancer.  The surgeon's responsibility is to maintain a covering of undisrupted normal tissue surrounding the malignancy.  
The surgeon must demonstrate continually that a crucial aspect of optimal gastrointestinal cancer surgery is adequate with intact lateral margins of dissection.  
Often a peritonectomy will facilitate a centripetal approach.(8)  
If the surgeon begins in deeper normal tissue planes he may dissect cleanly around the malignancy without disrupting the cancerous tissue.  
In centripetal surgery, one always performs the “easy” dissections first.  
One refuses to persist in any dissection where there is danger of hemorrhage, danger of damage to vital structures, or danger of developing an inadequate margin of normal tissue around the malignancy.

COMPLETE
LYMPHADENECTOMY

All gastrointestinal cancer operations require the resection of the lymphatic tissue that drains the cancer. The surgeon must perform the dissection which completely eliminates lymph node recurrence as a mechanism of failure.  The lymph nodes immediately adjacent to the intestine must be removed.  Also, all the intermediate and primary nodes down to the level of the celiac artery, superior mesenteric artery, or inferior mesenteric artery must be cleanly dissected away.  Because lymph node dissection adds very little, if at all, to the morbidity or mortality of the surgery, a maximal dissection including the complete second tier of lymph nodes is necessary.  A third tier of lymph nodes in and along the aorta and vena cava are thought to be past the point of diminishing return for lymphadenectomy.

There is a caveat with lymphadenectomy.  The surgeon should realize that it is naive to think that cancer exists within lymph nodes and that it is absent from lymphatic channels.  If lymphatic channels must be transected immediately adjacent to cancerous nodes, then one must assume that there has been intraoperative cancer spillage.

In summary, the ultimate goal that centripetal surgery allows the surgeon to pursue is cancer CONTAINMENT. 
Thorough gentle handling of the tissues surrounding the cancer, malignant cells are not disseminated into the remaining normal tissue.  Prevention of spillage of even one cancer cell into the operative site or into the peritoneal cavity must be a continuous concern to the surgeon who performs a gastrointestinal cancer surgery.  This containment combined with complete resection of lymph nodes draining the cancer is what constitutes adequate gastrointestinal cancer surgery.

REQUIREMENT FOR
INTRAPERITONEAL
CHEMOTHERAPY 

No matter how careful the surgeon is and how meticulous he is in his dissection, in some patients with advanced disease, there will be intraoperative cancer spill.  In this situation, perioperative intraperitoneal chemotherapy must be used.  This is the first indication for this technique listed in Tab. 1.  
The surgeon should employ heated intraoperative intraperitoneal chemotherapy using mitomycin C for the gastrointestinal adenocarcinomas.  The skin is tented up on a self-retaining retractor.  
Chemotherapy is maintained at 43oC within the peritoneal cavity.  The peritoneal perfusion must be carried on for at least an hour.  During the intraabdominal treatment, all of the intestines and other intraabdominal structures are continuously manipulated by the surgeon.  All residual blood clots and tissue debris must be washed away by the surgeon's hand because they are the matrix in which cancer cells progress (Fig. 1).

There are other absolute indications for heated intraoperative intraperitoneal chemotherapy.  If the surgeon removes lymph nodes containing cancer and these nodes are at the limits of the dissection, he must assume that cancer cells will be released unavoidably into the free peritoneal cavity. 
These spilled cancer cells will eventually cause the patient's death.  Intraperitoneal chemotherapy has been shown to be of benefit to gastric cancer patients with involved lymph nodes at the limits of dissection.(9)

Similarly, if in removing a malignancy exposed cancer tissue is seen at the margin of dissection, there is an extreme likelihood of contamination of the resection site.  
A chemotherapy wash of the peritoneal surfaces is necessary.

When the surgeon collects fluid from the abdominal or pelvic cavity, that fluid can be studied microscopically.  If cancer cells are seen then there is a positive peritoneal cytology. This patient is at extreme risk for cancer progression on abdominal or pelvic surfaces. 
This patient should have an intraperitoneal chemotherapy wash to eliminate microscopic residual disease.

If a gastrointestinal cancer perforates through the wall of the stomach or intestine, cancer cells have free access to the peritoneal surfaces.  If the cancer cells have established cancer nodules within the ovaries, this proves that there has been peritoneal cancer contamination.  
Likewise, if the cancer has grown all the way through the stomach or intestine to invade an adjacent organ or structure, peritoneal cancer contamination must be assumed.

Finally, in some instances, biopsy confirmed peritoneal seeding must be treated.  In these patients the selection factors for a palliative approach (debulking) as opposed to a curative approach (cytoreduction) have been clearly identified.(10)  Peritoneal seeding of limited distribution and limited mass should be curable in approximately 40% of patients if a combined treatment plan of peritonectomy procedures and perioperative intraperitoneal chemotherapy are used.

NOT ONLY THE PRIMARY
CANCER BUT ALSO 
MICROSCOPIC RESIDUAL
DISEASE 

How must surgeons change their attitude toward resection of gastrointestinal cancer?  The surgeon must consider himself responsible not only for resection of the large mass of primary cancer, but also for dealing with MICROSCOPIC RESIDUAL DISEASE.  
Sometimes he does this by using centripetal surgery to prevent the spread of cancer cells.  
At other times he does it by treating carcinomatosis using peritonectomy procedures combined with peritoneal perfusion using a chemotherapy solution.  

There are some important consequences of this view of surgical skill and cancer survivorship.  
Obviously, if there are such widely divergent results, then some surgeons are doing it wrong.  
Surgeons must demonstrate their expertise by maintaining a personal scorecard.  
Surgeons who have a high local recurrence rates and poor survival rates must be eliminated from the credentialed group.  
Patients with gastrointestinal cancer must be protected from needless death.

Acceptance of this concept means that approximately 30% of gastric cancer patients and 10% of colon cancer patients who present with a primary cancer with peritoneal seeding should be treated in specialized centers where the surgeons are familiar with peritonectomy procedures and  intraoperative intraperitoneal chemotherapy  treatment is readily available.  
The results of treatment of peritoneal carcinomatosis are far superior when the primary tumor and the involved peritoneum are removed simultaneously.(10,11)  
At the same operative setting, chemotherapy should be used to eliminate microscopic residual disease from the cancer resection site and from other places on the peritoneal surface.  
If the primary cancer with peritoneal seeding is resected in the absence of a chemotherapy wash, the cancer cells on the peritoneal surface will be implanted deep within the tissues and will be difficult, if not impossible to eradicate.

What is the expected outcome if surgeons prevent or treat microscopic residual disease?  If the requirements presented here are fulfilled the survival rate with gastric cancer should increase to approximately 65% overall.  Currently, in the United States it is approximately 30%.  
The survivorship of colon and rectal cancer would increase to 90%.  Currently it is approximately 45%.  It is likely that surgical skill is a more important determinant of prognosis than the aggressive nature of the cancer or its stage at diagnosis.

In conclusion, peritoneal dissemination can be prevented by the responsible surgeon at least in part by proper surgical technique which is utilized to resect the primary malignancy.  
Containment must be the number one priority of the gastrointestinal cancer surgery.  
Also, established peritoneal carcinomatosis can be cured if it is attacked in a timely fashion with heated intraoperative intraperitoneal chemotherapy, peritonectomy procedures, and visceral resections for cancer on bowel surfaces.  
Many little changes can make a big difference in survival with gastrointestinal cancer surgery.  
Surgery must do all that it can to benefit the patient with the facilities and time that one has available with acceptable morbidity and mortality.  
Adequate surgical skill implies that cancer spread is prevented or treated and thereby optimizes survivorship.
 

Paul H. Sugarbaker
Director, Surgical Oncology
Peritoneal Carcinomatosis Program
Center for Surgical Oncology
Washington Hospital Center