Lawrence J. Burgart

The Frequency of Hereditary Defective Mismatch Repair in a Prospective Series of Unselected Colorectal Carcinomas

A comprehensive analysis of somatic and germline mutations related to DNA mismatch-repair (MMR) genes can clarify the prevalence and mechanism of inactivation in colorectal carcinoma (CRC). In the present study, 257 unselected patients referred for CRC resection were examined for evidence of defective DNA MMR. In particular, we sought to determine the frequency of hereditary defects in DNA MMR in this cohort of patients. MMR status was assessed by testing of tumors for the presence or absence of hMLH1, hMSH2, and hMSH6 protein expression and for microsatellite instability (MSI). Of the 257 patients, 51 (20%) had evidence of defective MMR, demonstrating high levels of MSI (MSI-H) and an absence of either hMLH1 (n=48) or hMSH2 (n=3). All three patients lacking hMSH2, as well as one patient lacking hMLH1, also demonstrated an absence of hMSH6. DNA sequence analysis of the 51 patients with defective MMR revealed seven germline mutations-four in hMLH1 (two truncating and two missense) and three in hMSH2 (all truncating). A detailed family history was available for 225 of the 257 patients. Of the seven patients with germline mutations, only three had family histories consistent with hereditary nonpolyposis colorectal cancer. Of the remaining patients who had tumors with defective MMR, eight had somatic mutations in hMLH1. In addition, hypermethylation of the hMLH1 gene promoter was present in 37 (88%) of the 42 hMLH1-negative cases available for study and in all MSI-H tumors that showed loss of hMLH1 expression but no detectable hMLH1 mutations. Our results suggest that, although defective DNA MMR occurs in approximately 20% of unselected patients presenting for CRC resection, hereditary CRC due to mutations in the MMR pathway account for only a small proportion of patients. Of the 257 patients, only 5 (1.9%) appear to have unequivocal evidence of hereditary defects in MMR. The epigenetic (nonhereditary) mechanism of hMLH1 promoter hypermethylation appears to be responsible for the majority of the remaining patients whose tumors are characterized by defective DNA MMR.

Inferior mesenteric venous sampling to detect colonic ischemia: A comparison withlaser Doppler flowmetry and photoplethysmography

PURPOSE No single method has been identified that accurately and reliably detects patients with impending bowel infarction during aortic reconstruction. Serial sampling of blood gas from the inferior mesenteric vein (IMV) for detecting colonic ischemia was compared with two previously described techniques: laser Doppler flowmetry (LDF) and photoplethysmography. METHODS Nine dogs underwent induced partial colonic ischemia followed by complete ischemia. Serial IMV blood gas measurements were obtained at four intervals: baseline, partial ischemia, complete ischemia, and reperfusion. Simultaneous direct colon wall LDF and PPG measurements also were obtained. RESULTS Changes in pH, Po2, O2 saturation, and Pco2 demonstrated progressive acidosis, hypoxemia, and hypercapnia in association with progressive arterial occlusion and a reversal of these trends toward baseline after restoration of flow. The absence of a pulsatile photoplethysmography tracing and oxygen saturation less than 90% were predictive of altered perfusion but could not differentiate partial from complete ischemia. Although the differences in mean LDF values were statistically different during ischemia and reperfusion, there was considerable variability between each measurement. CONCLUSIONS Analysis of blood gas from the IMV and pulse oximetry are useful techniques for detecting colonic ischemia, but only the former can distinguish partial from complete ischemia. The variability in colonic measurements with LDF limits its usefulness for detecting levels of colonic perfusion.

3637 Efficacy of submucosal saline injection in the limitation of colonic thermal injury by electrosurgical devices.

Many colonoscopic therapies involve cutting or coagulation effects via delivery of thermal injury. Complications of these techniques include symptomatic transmural burn (post-polypectomy syndrome) and perforation. Submucosal saline injection (SMSI) has been advocated as a means of limiting depth of injury. The saline cushion acts as a heat-sink and increases transmural distance from burn to serosa. However, data are lacking regarding the efficacy of SMSI in limiting depth of thermal injury. Aim: To determine, in a porcine model, the effect of SMSI on depth of thermal injury to the colon due to a various modalities. Methods: Laparotomy under general anesthetic was performed and a longitudinal colotomy incision made on the antimesenteric side. Burns were made using bipolar goldprobe (20W, 2sec), heater probe (HP; 30J); monopolar (MP) contact with biopsy forceps (20W, 2sec), and MP non-contact with argon plasma coagulation (APC; 45W, 3sec); n ≥11 for each lesion. Burns were with or without prior injection of 2mL saline. The incision was closed and animals killed at 24hrs. Lesions were excised for histologic analysis. Injury was assessed by severe damage to the deep (longitudinal) muscle layer. Results: Non-SMSI lesions resulted in deep muscle injury in 86%, 61%, 50% and 18% for APC, MP contact, HP & bipolar, respectively. SMSI reduced risk of deep injury for APC and HP, but not monopolar contact (86%→21%; 50%→0% & 61%→50%, respectively). Conclusions: At equivalent energy outputs, bipolar current results in less deep injury than MP current or HP. Prolonged coagulation with the APC (45W) results in deep colonic injury. At the settings used, saline injection limits depth of injury due to both heater probe and APC, but not monopolar contact cautery. In spite of SMSI, caution should be used with prolonged monopolar cautery.

3470 Widespread circumferential mucosectomy in the porcine esophagus–a pilot study.

Endoscopic mucosal resection (EMR) is a procedure growing in significance as an alternative to surgery in the treatment of gastrointestinal mucosal lesions. Current EMR techniques are limited to lesions measuring less than 2 cm. Aim: To demonstrate that widespread circumferential mucosectomy up to 5 cm in length can be performed using a stripping method. Method: Eight, 60 kg anesthetized pigs were studied. The distal 5 cms of the esophagus was injected with 50% dextrose to create a lasting submucosal protective cushion. A strip of mucosa approximating 5 cm in length from the esophagogastric junction and 1 cm in width was marginated along its length and proximal end using a porcelain ball-tipped needle knife and prototype teflon backed monopolar scalpel (Olympus America, Inc., NY). The proximal end once undermined was grasped using forceps and stripped off the submucosal cushion in a distal direction. The distal end was then resected using a needle knife from a retroflexed position. This was repeated sequentially until the distal esophagus was circumferentially denuded of mucosa. Four pigs were sacrificed immediately after the procedure while two each of the remaining four animals were followed for two and four weeks. Result: Complete widespread circumferential mucosectomy was achieved in five animals with an average of four strips removed per pig. The average strip length was 5.5 cm(range 5-6 cm) and width 0.9 cm(range 0.6-1.2 cm). There were no procedural complications in the five animals. Mucosectomy was incomplete in the first two animals as the technique of mucosal stripping evolved with another animal experiencing a perforation early in this learning curve. The animals sacrificed immediately post-procedure showed histologically intact muscularis propria with some preservation of the submucosa and minimal associated edema and cautery injury. Repeat endoscopy of the surviving animals (4) showed healing with moderate (1) to tight (3) strictures at both two and four weeks. Conclusion: 1. Widespread circumferential esophageal mucosectomy can be performed endoscopically using a true stripping technique with preservation of the muscularis propria. 2. This technology has exciting potential clinical applications especially in the treatment of Barretts esophagus. 3. The appropriate measures to prevent post-procedural stricture formation need to be further investigated and identified. (The Apollo Group: Sydney Chung, Peter Cotton, Christopher Gostout, Robert Hawes, Anthony Kalloo, Pankaj Pasricha, Thadeus Trus)

7001 Digital image analysis (dia) improves the diagnostic yield of endoscopic biliary brush cytology.

DIA allows quantification of nuclear DNA content that may aid in distinguishing benign from malignant biliary tract strictures. Methods: The Mayo Clinic pathology databank was used to identify all biliary brush cytology specimens obtained between 6/97 and 6/99. Corresponding medical records were reviewed to determine whether patients had malignant or benign biliary strictures. A stricture was classified as malignant if proven by biopsy or surgery, and as benign if confirmed by surgery or a cancer-free clinical course for ≥ 1 year. Strictures were further classified into benign strictures with negative brush cytology (Group 1), malignant strictures with negative brush cytology (Group 2a- pancreatic carcinoma, Group 2b- cholangiocarcinoma), and malignant strictures with positive brush cytology (Group 3). A significant number of specimens were not available for analysis, as they were prepared on filters. Pap-stained smears of available brush cytology specimens were de-stained and then re-stained using Feulgen dye. A maximum of 200 nuclear images were quantified for DNA content without knowledge of stricture type. DNA histograms were generated and ploidy results compared with the class of stricture. Results: 25 specimens were analyzed (see table). Assuming the presence of any aneuploid cells indicating malignancy, the sensitivity of DIA was 85%, accurately predicting a malignant stricture in 13 of 16 samples. Conclusions: 1. Ploidy assessment by DIA has the potential to greatly enhance the sensitivity of diagnosing malignant strictures compared to routine cytology alone. 2. Additional studies will be needed to further assess the specificity of DIA in a broad range of benign inflammatory lesions.