Subcutaneous rather than intravenous ustekinumab induction is associated with comparable circulating drug levels and early clinical response: a pilot study.

BACKGROUND: Ustekinumab (USK) is licenced for intravenous induction and subcutaneous (S/C) maintenance in Crohn's disease. AIM: To evaluate ustekinumab trough concentrations and clinical response with exclusive subcutaneous ustekinumab induction. METHODS: Patients with Crohn's disease who initiated treatment with subcutaneous ustekinumab at a single academic centre were included in this pilot study. A dosage of 360 mg ustekinumab was given subcutaneously in divided doses; 180 mg at Week 0, 90 mg at Week 1 and 90 mg at Week 2, with corresponding ustekinumab trough concentrations assessed to Week 8. The primary outcome measures were trough serum ustekinumab levels and clinical remission at Week 8. Secondary outcome measures were trough serum ustekinumab levels at Week 1 & 2 and changes in C-reactive protein, albumin and faecal calprotectin at Week 8. RESULTS: Nineteen patients were included. Median Week 8 ustekinumab trough concentrations were 6.1 µg/mL (Inter-quartile range 4-9.8 µg/mL). There was a significant improvement in Harvey Bradshaw index from Week 0 (median HBI 5; interquartile range 2-8) to Week 8 (median HBI 1; interquartile range 0-3) (P = 0.002). C-reactive protein levels did not change significantly but faecal calprotectin improved significantly; median faecal calprotectin at Week 0 was 533 µg/g; at Week 8, it was 278 µg/g (P = 0.038). CONCLUSIONS: Ustekinumab trough concentrations are comparable whether ustekinumab induction treatment was administered subcutaneously or intravenously. A significant improvement in symptoms and faecal calprotectin was noted. These results support the use of subcutaneous induction as an alternative if there are barriers to intravenous induction.

Spatial variation of the colonic microbiota in patients with ulcerative colitis and control volunteers.

OBJECTIVES: The relevance of spatial composition in the microbial changes associated with UC is unclear. We coupled luminal brush samples, mucosal biopsies and laser capture microdissection with deep sequencing of the gut microbiota to develop an integrated spatial assessment of the microbial community in controls and UC. DESIGN: A total of 98 samples were sequenced to a mean depth of 31,642 reads from nine individuals, four control volunteers undergoing routine colonoscopy and five patients undergoing surgical colectomy for medically-refractory UC. Samples were retrieved at four colorectal locations, incorporating the luminal microbiota, mucus gel layer and whole mucosal biopsies. RESULTS: Interpersonal variability accounted for approximately half of the total variance. Surprisingly, within individuals, asymmetric Eigenvector map analysis demonstrated differentiation between the luminal and mucus gel microbiota, in both controls and UC, with no differentiation between colorectal regions. At a taxonomic level, differentiation was evident between both cohorts, as well as between the luminal and mucosal compartments, with a small group of taxa uniquely discriminating the luminal and mucosal microbiota in colitis. There was no correlation between regional inflammation and a breakdown in this spatial differentiation or bacterial diversity. CONCLUSIONS: Our study demonstrates a conserved spatial structure to the colonic microbiota, differentiating the luminal and mucosal communities, within the context of marked interpersonal variability. While elements of this structure overlap between UC and control volunteers, there are differences between the two groups, both in terms of the overall taxonomic composition and how spatial structure is ascribable to distinct taxa.

Tumour microenvironment of both early- and late-stage colorectal cancer is equally immunosuppressive.

BACKGROUND: Tumour microenvironment (TME) of advanced colorectal cancer (CRC) suppresses dendritic cell (DC) maturation. Here, our aim was to determine how the microenvironment of early-stage tumours influences DCs. METHODS: Tumour-conditioned media (TCM) was generated by culturing explant tumour tissue in vitro (n=50). Monocyte-derived DCs (MDDCs) of healthy donors or cancer patients were pretreated with TCM and stimulated with lipopolysaccharide (LPS). DC maturation was assessed by flow cytometry and cytokine production measured by ELISA. RESULTS: TCM from both early- and late-staged tumours abrogated LPS-induction of IL-12p70 secretion, while increasing IL-10. The profile of inflammatory mediators in TCM was similar across stages, and all increased pSTAT3 expression by DCs.CRC patient DCs (n=31) secreted low levels of IL-12p70 and failed to upregulate expression of maturation markers in response to LPS. Furthermore, in vitro culture of autologous DCs with TCM did not change the hypo-responsiveness of patient DCs. CONCLUSION: Our data demonstrates that the TME of all stages of CRC contains inflammatory mediators capable of suppressing local DCs. MDDCs obtained from CRC patients are hyporesponsive to stimuli such as LPS. Measures to reverse the negative influence of the TME on DCs will optimise cancer vaccines in both early- and late-stage CRC.

Infliximab therapy in Crohn's disease: a pragmatic approach?

BACKGROUND: Infliximab is recognized as an effective therapy in unresponsive luminal and fistulating Crohn's disease. The use of maintenance or 'on demand' therapy thereafter is controversial. AIM: To assess the need for maintenance infliximab therapy in a clinical setting where immunomodulatory agents are widely used and where episodic therapy is used in preference to maintenance therapy. METHODS: Ninety-three patients with Crohn's disease receiving infliximab; 72 with unresponsive luminal disease and 21 with fistulous disease. Data collected included disease site and duration, surgical and smoking history, initial response rates, duration of response maintenance and concomitant medications. RESULTS: Fifty-six of 72 (78%) patients with luminal disease and 11 of 21 (52%) with fistulous disease achieved an initial response. Ten of 67 responders required conversion to maintenance infliximab infusions, while 31 remain in remission. Patients with luminal disease and those who had not taken previous surgery had higher response rates to infliximab. Younger patients and those with small bowel disease had higher relapse rates following initial response. Three patients developed allergic reactions to infliximab and one patient died of progressive pulmonary disease 6 weeks after their first infusion. CONCLUSIONS: Many patients with Crohn's disease can be maintained successfully with an episodic infliximab regimen.

Direct-vision stenting: the way forward for malignant oesophageal obstruction.

BACKGROUND AND STUDY AIMS: Self-expanding metallic stents now form the mainstay of treatment for palliation of dysphagia in oesophageal cancer. These stents are generally inserted under fluoroscopic guidance. However, both the internal and external marking of the tumour can be inaccurate and time-consuming, and access to fluoroscopic facilities is sometimes limited. We prospectively assessed the use of a method of stent insertion under direct vision without the aid of fluoroscopy. PATIENTS AND METHODS: A total of 50 consecutive patients presenting with obstructive symptoms secondary to inoperable oesophageal cancers were included in the study. We used either the 7-cm or the 11-cm covered Choo stent (MI-Tech Ltd., Seoul, South Korea). RESULTS: A total of 52 stents were inserted under direct vision. The procedure generally took less than 15 minutes and good palliation was achieved without complications. Fluoroscopic assistance was required in only one patient. CONCLUSIONS: Direct-vision stent insertion is simple, safe, effective, and only rarely requires fluoroscopic assistance. The technique may be of particular use in centres with limited access to fluoroscopy.

Pathological response following long-course neoadjuvant chemoradiotherapy for locally advanced rectal cancer.

AIMS: To standardize the pathological analysis of total mesorectal excision specimens of rectal cancer following neoadjuvant chemoradiotherapy for locally advanced disease (T3/T4), including tumour regression. METHODS AND RESULTS: Standardized dissection and reporting was used for 60 patients who underwent total mesorectal excision following long-course chemoradiotherapy. Tumour regression was scored by two pathologists (K.S., D.G.) using both an established 5-point tumour regression grade (TRG), and a novel 3-point grade. Both scores were evaluated for interobserver variability. A complete or near-complete pathological response (3-point TRG 1) was found in 10 patients (17%). Using the 5-point TRG, there was good agreement between both pathologists (kappa = 0.64). Using the 3-point grade, agreement was excellent (kappa = 0.84). No disease recurrence has been reported in patients with a complete, or near complete pathological response (3-point TRG 1), after a mean follow-up of 22 months. CONCLUSION: Tumour regression grade is a useful method of scoring tumour response to chemoradiotherapy in rectal cancer. TRG 1 and 2 can be regarded as a complete pathological response (ypT0). A modified 3-point grade has the advantage of better reproducibility, with similar prognostic significance.

Expression of Cathepsin B and L antigen and activity is associated with early colorectal cancer progression.

Cathepsin B and Cathepsin L are cysteine proteases important in the process of invasion and metastasis. The aim of our study was to assay antigen and activity levels of these enzymes and to correlate these with established clinical and pathological prognostic parameters including patient survival. 99 patients undergoing operations for colorectal cancer were included in this study. We quantitated cathepsin B and L levels in matched normal mucosa and cancer samples using an enzyme-linked immunosorbent assay (ELISA) and specific activity assays and expressed the results as tumour/normal ratios. Significant correlations were found between tumour/normal cathepsin B and L antigen and activity ratios. Cathepsin B and L tumour/normal activity ratios were greater than 1 in early stage disease and there were gradual reductions in cathepsin B (P = 0.02) and L (P = 0.03) activity ratios with advancing tumour stage. Survival of patients with potentially curative disease was inversely related to both cathepsin B (P = 0.007) and L (P = 0.001) activity ratio, in addition to cathepsin L antigen ratio (P = 0.008). Our findings suggest that cysteine proteases play an important role in colorectal cancer progression.

Upper gastrointestinal endoscopy performed by nurses: scope for the future?

BACKGROUND: Previous researchers have shown that non-medical endoscopists can perform lower gastrointestinal endoscopy as safely and effectively as medical staff. However, it is not known if upper gastrointestinal endoscopy performed by medical and non-medical endoscopists in clinical practice yields similar results in terms of performance, patient discomfort, and satisfaction. AIM: To determine differences in the yield of diagnosis for significant disease during upper gastrointestinal endoscopy performed by nurse and medical endoscopists and to measure patient discomfort, satisfaction, and attitudes towards future endoscopy. PATIENTS: This two part study included 3009 patients in a retrospective analysis and 480 in a prospective study. METHODS: The first part of the study assessed indications for endoscopy, diagnoses, and procedures performed by medical and nurse endoscopists. In a second prospective study, 480 patients were included to determine the association between endoscopist type and sedation, patient anxiety, discomfort, satisfaction, and attitudes towards future sedation. RESULTS: No patient refused endoscopy by either a nurse or medical endoscopist and there were no complications in either group. Nurses performed 1487 procedures and reported fewer endoscopies as "normal" than medical staff (p=0.006). Multivariate analysis showed that male sex, older age, inpatient status, dysphagia, and gastrointestinal bleeding, but not endoscopist type, were all associated with significant disease. In relation to discomfort and satisfaction, a similar proportion of patients received sedation in both groups (p=0.81). There were no differences in pre-procedure anxiety (p=0.61), discomfort during intubation (p=0.97), discomfort during examination (p=0.90), or post-procedure examination rating (p=0.79) in patients examined by medical or nurse endoscopists. CONCLUSION: Experienced nurses perform routine diagnostic gastroscopy safely in everyday clinical practice and with as little discomfort and as much patient satisfaction as medical staff.

Peritoneal involvement in stage II colon cancer.

A pathologist (K.S.) reviewed histologic slides for peritoneal involvement by tumor cells for 118 patients with stage II colon cancer. Patients were followed up for a median of 6 years. Tumor cells were found free in the peritoneal space in 16 cases (13.6%). The presence of cancer cells free in the peritoneal space was associated with lymphovascular invasion (P = .001) and neural invasion (P < .001). The overall 5-year survival was 80% in the patient population, but was 39% and 86% for those with and without tumor cells free in the peritoneal space, respectively (P < .0001). Multivariate analysis confirmed that free tumor cells within the peritoneal space (P < .0001) and lymphovascular invasion (P = .007) were related independently to outcome. Peritoneal involvement with tumor cells free in the peritoneal space in stage II colon cancer is a powerful indicator of outcome; patients have a survival similar to that for patients with stage III disease.

Yield of colonoscopy in patients with nonacute rectal bleeding: a multicenter database study of 1766 patients.

OBJECTIVE: There are few data to guide the choice between colonoscopy and flexible sigmoidoscopy in patients with nonacute rectal bleeding, especially in younger age groups. Our aim was to determine the yield of colonoscopy for significant proximal large bowel disease in the absence of significant distal disease, with special reference to young patients. METHODS: This was a retrospective study of data collected prospectively in 1766 patients (median age 57 yr, 711 women). The endoscopic database (GI-Trac) contained 152 discrete fields for data input. Multiple logistic regression analysis was performed to identify variables independently associated with the presence of isolated significant proximal disease. RESULTS: Young patients had a higher percentage of normal examinations than did older patients. The incidence of diverticular disease, small polyps, large polyps, and cancer rose with increasing age. No patient aged <40 yr had an isolated proximal cancer, but 7% had other significant isolated proximal disease. There was no overall association between age and significant proximal disease in the absence of significant distal disease (p = 0.66). The only variable associated with isolated proximal disease was anemia (odds ratio = 1.81; 95% CI = 1.11-2.93; p = 0.02). CONCLUSION: The yield of colonoscopy (beyond the range of sigmoidoscopy) for neoplasia is low in patients aged <40 yr, but other significant disease may be missed if age is the only criterion determining colonoscopy use.

Alu repeat sequences are present in increased proportions compared to a unique gene in plasma/serum DNA: evidence for a preferential release from viable cells?

Small amounts of DNA circulate freely in plasma or serum, but the mechanism of release is not known. To determine if DNA is actively excreted from viable cells, we utilized real-time PCR to measure the proportion of Alu repeat sequences compared to the beta-globin gene in serum and lymphocyte DNA in 27 cancer patients and 22 healthy controls. The proportion of Alu compared to beta-globin was significantly greater in serum DNA than in lymphocyte DNA both in control subjects (p = 0.003) and in cancer patients (p < 0.001). Overall, the proportion was similar in cancer and control patients (p = 0.79). Further experiments showed that the beta-globin gene was not more vulnerable to degradation by nuclease action than Alu sequences. Our results lead us to conclude that active DNA release is likely to play a significant role in the origin of circulating DNA.

A prospective controlled trial of an ultrathin versus a conventional endoscope in unsedated upper gastrointestinal endoscopy.

BACKGROUND AND STUDY AIMS: Upper gastrointestinal endoscopy is performed without sedation in many countries. Unsedated patients experience more discomfort during endoscopy than sedated patients, but few studies have examined factors which could be modified to minimize discomfort during the procedure. We assessed the effect of endoscope diameter on patient discomfort during unsedated transoral gastroscopy. PATIENTS AND METHODS: A total of 322 patients attending for unsedated endoscopy were examined using an endoscope of diameter either 6.0 mm or 9.8 mm. Patients completed a two-part questionnaire assessing tolerance of the procedure and discomfort during it. RESULTS: There was failure to complete the initial unsedated endoscopy in three of 163 patients in the 6.0 mm group and 14 of 159 in the 9.8 mm group (P = 0.009). Patients in the 6.0 mm group reported less discomfort both during endoscope insertion (P < 0.0001) and during the remainder of the procedure (P < 0.0001). 14% of patients in the 6.0 mm group indicated that they would request sedation if a further endoscopy were necessary, compared with 31% in the 9.8 mm group (P = 0.0005). CONCLUSIONS: Ultrathin endoscopes may have a role in clinical practice if randomized comparative studies with standard-bore instruments confirm that they do not compromise diagnostic quality.

Pulse oximetry saturation levels during routine unsedated diagnostic upper gastrointestinal endoscopy.

BACKGROUND: Diagnostic procedures account for over 90% of all upper gastrointestinal endoscopies. Pulse oximetry saturation (SpO2) levels were assessed in patients attending for routine unsedated diagnostic gastroscopy to identify factors associated with oxygen desaturation. METHODS: Three hundred and thirty unsedated patients were monitored with continuous pulse oximetry. A further 154 patients who requested sedation prior to endoscopy were studied as a positive control group. RESULTS: SpO2 levels were lower in sedated compared to unsedated patients (P < 0.0001). Six unsedated patients (2%) desaturated to 90% or less during endoscopy compared to 32 sedated patients (21%) (P < 0.0001). SpO2 levels in unsedated patients were not related to patient sex, age, cigarette smoking, endoscope diameter, basal SpO2 levels or duration of endoscopy. In contrast, examination of the pharyngeal area and epiglottis (P = 0.0002) and a longer intubation time (P = 0.0002) were associated with lower SpO2 levels. The lowest SpO2 level recorded during unsedated endoscopy was 88%. CONCLUSION: Pulse oximetry is not a prerequisite to performing routine unsedated diagnostic gastroscopy in patients without severe systemic disease.

Changing patterns of sedation use for routine out-patient diagnostic gastroscopy between 1989 and 1998.

BACKGROUND: Knowledge of sedation trends for upper gastrointestinal endoscopy is important for health service planning, particularly in view of rapidly increasing demands on endoscopy services. However, no data are available on sedation trends in Britain over the past 10 years. AIM: To determine sedation use for routine gastroscopy in a single endoscopy unit between 1989 and 1998. METHODS: This was a retrospective study of 9795 consecutive adults (mean age 56 years, range 18-100 years; 4512 females) who had undergone a gastroscopy between 1989 and 1998. Clinical, pharmacological and endoscopic data were retrieved from a computerized database. RESULTS: Over the 10-year study period, the sedation rate remained constant for patients undergoing therapeutic endoscopy (P=0.99) and those undergoing in-patient diagnostic examinations (P=0.63). In contrast, the sedation rate for out-patient diagnostic endoscopy decreased by 54%, from a high of 70% in 1990 to 32% in 1998 (P < 0.0001). Logistic regression analysis showed that the decline in sedation use was greater in females (P < 0.0001) than males and in procedures performed by non-consultant compared to consultant staff (P=0.01). CONCLUSIONS: If our results form part of a national trend, they will have important implications for cardiopulmonary monitoring strategies, recovery room practices and for complication rates due to the use of sedation for upper gastrointestinal endoscopy.

Factors associated with tolerance to, and discomfort with, unsedated diagnostic gastroscopy.

BACKGROUND: Identification of patients likely to experience high levels of discomfort during unsedated gastroscopy would be useful as these patients could be prospectively targeted for sedation. We prospectively assessed patient and endoscopic variables in subjects attending for endoscopy in order to identify factors associated with patients' experience of the unsedated examination. METHODS: We studied 508 patients attending for routine diagnostic gastroscopy. Clinical and endoscopic data were collected and patients completed a two-part questionnaire assessing their anxiety with, and experience of, the procedure. RESULTS: Thirty-nine subjects failed to complete the initial unsedated endoscopy. Failure to tolerate endoscopy was associated with younger age (P = 0.002) and examination with a standard-bore (> or = 9.0 mm) endoscope (P = 0.004). High levels of patient discomfort during the procedure were associated with younger age (P < 0.001), high levels of pre-endoscopic anxiety (P < 0.001), high levels of pre-endoscopic discomfort due to throat spray (P = 0.02) and examination with a standard-bore endoscope (P < 0.001). Preference for sedation during future examinations was related to female gender (P = 0.02), young age (P = 0.02), high levels of apprehension (P < 0.001), the examining doctor (P = 0.002) and use of a standard-bore endoscope (P < 0.001). CONCLUSIONS: Discrete clinical characteristics and endoscopic variables are associated with patients' experience of unsedated endoscopy. Further work might result in an algorithm for identifying patients who would benefit from sedation prior to gastroscopy.

Telomerase RNA as a detection marker in the serum of breast cancer patients.

Tumor-derived circulating DNA has been found in the plasma of cancer patients. Alterations include decreased strand stability, mutations of oncogenes or of tumor suppressor genes, microsatellite alterations, and hypermethylation of several genes. RNA has also been found circulating in the plasma of normal subjects and cancer patients. Tyrosinase mRNA has been extracted from the serum of melanoma patients and subjected to RT-PCR. Moreover, the presence of cell-free EBV-associated RNA has been reported in the plasma of patients with nasopharyngeal carcinoma. Human telomerase comprises two RNA subunits, telomerase RNA template (hTR) and its catalytic component, telomerase reverse transcriptase protein (hTERT). Expression of these subunits correlates with telomerase activity. Using RT-PCR, we investigated whether these RNA subunits were present in the serum of 18 patients with breast cancer, 2 patients with benign breast disease, and 21 normal subjects. The presence of amplifiable RNA was confirmed in all tissue and serum samples using RT-PCR of glyceraldehyde-3-phosphate dehydrogenase RNA. hTR was found in 17 of 18 tumors (94%) and 5 of 18 serum samples (28%). hTERT was also detected in 17 of 18 tumors (94%) and in 4 of 16 available serum samples (25%). hTR and hTERT were undetectable in tissues and sera taken from 2 patients with benign disease and in the sera of 21 normal subjects. We conclude that RNA is detectable in the serum of breast cancer patients and that tumor-derived mRNA can be extracted and amplified using RT-PCR, even in patients with localized disease. This may have implications for cancer diagnosis and follow-up in the future.

p27(Kip1) loss does not predict survival in patients with advanced gastric carcinoma.

BACKGROUND: p27(Kip1) is a cyclin-dependent kinase inhibitor whose loss is associated with disease progression and an unfavorable outcome in several malignancies. The authors studied its expression in a consecutive series of resected gastric carcinomas. METHODS: Expression of p27(Kip1) in 71 advanced gastric carcinomas and 10 lymph nodes containing metastases was determined using an avidin-biotin-peroxidase immunohistochemical method. The relations between p27(Kip1) expression and pathologic features, patient characteristics, and survival were analyzed. RESULTS: p27(Kip1) levels in gastric carcinomas ranged from 0.63-82.97% (median, 23. 10%; mean, 27.99%). There was no association found between p27(Kip1) expression and patient gender (P = 0.21), patient age (P = 0.13), tumor stage (P = 0.17), tumor grade (P = 0.22), or histologic type (P = 0.72). Univariate analysis showed that long term survival was related to stage (P < 0.0001) and grade (P = 0.03). However, tumors with p27(Kip1) levels above and below the median value were associated with a similar outcome, regardless of whether all cases (P = 0.19) or those without metastatic disease (P = 0.50) or those with residual or metastatic disease (P = 0.92) were included. When entered into a multivariate analysis, stage (P < 0.0001) and grade (P = 0.05), but not p27(Kip1) levels (P = 0.16), were found to be related to patient outcome. In lymph node metastases, p27(Kip1) expression (median, 16.5%) was similar to that found in the corresponding primary lesion (median, 30.9%). CONCLUSIONS: p27(Kip1) may play a role in the pathogenesis and progression of gastric carcinoma, but its expression is unlikely to be useful as a prognostic indicator, at least in European patients with advanced disease.

bcl-2 protein expression is associated with better prognosis in colorectal cancer.

AIMS: To investigate the expression of bcl-2 in colorectal carcinoma and to examine its association with mediators of apoptosis (p53 and mdm-2), clinicopathological features and long-term outcome. METHODS AND RESULTS: We determined by immunohistochemistry the expression of bcl-2 in 102 colorectal carcinomas with 10-year follow-up. In 66 of these cases in which we had previously assessed p53 status, no correlation was seen between bcl-2 and p53. The mdm-2 protein was not detected in any of the 66 cases. Cytoplasmic staining of the bcl-2 gene product was seen in the tumour cells of 22 cases (22%). Using a polymerase chain reaction technique we showed that overexpression of bcl-2 was not due to rearrangement of the bcl-2 gene. Expression of bcl-2 protein was related to tumour grade but was unrelated to patient age, sex, tumour site, tumour size or Dukes' stage. There was a trend towards increased survival in those whose tumours expressed bcl-2 protein (P = 0.055). When entered into a multivariate analysis, this survival difference was independent of tumour stage (P = 0.05). CONCLUSIONS: These results suggest that bcl-2 expression in colorectal carcinoma is associated with a better long-term prognosis.

p53 expression in phyllodes tumours is associated with histological features of malignancy but does not predict outcome.

AIMS: To study p53 protein expression in phyllodes tumours of the breast, with particular attention to its prevalence and to its relationship with histological features and clinical outcome. METHODS AND RESULTS: Stromal and epithelial p53 immunohistochemical expression was studied in 57 phyllodes tumours (27 benign, 17 borderline, 13 malignant) using an avidin-biotin peroxidase method. High levels of expression (> 30% of stromal nuclei) were found in eight phyllodes tumours (14%). p53 expression was associated with tumour grade (P = 0.001), prominent stromal overgrowth (P = 0.0003), prominent stromal nuclear pleomorphism (P = 0.006), high stromal mitotic count (P = 0.05), and an infiltrative tumour margin (P = 0. 05). Six patients were lost to follow-up after surgery. Mean follow-up time of the remaining 51 patients was 7.3 years (median 4. 3, range 0.5-25) or until death. Sixteen patients (31%) experienced tumour recurrence. Recurrence was more likely if there was an infiltrative tumour margin (P = 0.006) or prominent stromal overgrowth (P = 0.04) but not p53 expression (P = 0.55). A minority of recurrences expressed p53 more extensively than their primary counterparts. There were five tumour-related deaths (10% of patients). Death was associated with high grade (P = 0.0002), prominent stromal overgrowth (P = 0.0001), an infiltrative margin (P = 0.0002), prominent nuclear pleomorphism (P = 0.005), a high mitotic count (P = 0.01) and tumour size (P = 0.03). Again, p53 expression was not associated with tumour-related survival (P = 0. 13). CONCLUSIONS: p53 abnormalities occur in a minority of borderline and malignant phyllodes tumours. p53 expression is associated with known negative prognostic factors, but does not appear to be a useful determinant of tumour recurrence or long-term survival.