A total of 93 patients with MPM were included in this study. The specimens were obtained from patients who underwent surgical resection at the Department of Cardio-Thoracic Surgery of the University of Pisa, from January 1999 to May 2008. Participation in this study required informed consent. No patient had received chemotherapy or radiotherapy before surgery. The patients were divided into two groups: patients who received treatment with preoperative intra-pleural IL-2 and untreated patients. Sixty patients, 51 males and 9 females received preoperative immunotherapy with intra-pleural IL-2 (18 × 106 UI per day every other day for three times), according to eligibility criteria and the treatment protocol described earlier (Lucchi et al, 2007). After 1 day of recovery, the patients underwent a thoracotomy. Of the 60 patients, 53 (88.3%) had a pleurectomy/decortication (P/D), whereas the remaining 7 patients (11.7%) had an extrapleural pneumonectomy. Thirty-three patients, 27 men and 6 women underwent surgical resection without preoperative immunotherapy with IL-2 or any other preoperative therapy. All of the 33 patients had a P/D.

Sixty patients were observed until death or the final date of analysis (August 2008), with a median follow-up time of 74 months for living patients (range, 2–113 months). Forty-one (68.3%) patients relapsed during follow-up: 33 had a local relapse, 7 had both a local and a systemic relapse, and only 1 patient had a systemic relapse. At the end of the follow-up period, 15 patients (25%) were still alive and 45 (75%) had died. Time to progression and overall survival rates were calculated as the period from surgery until the date the disease started to worsen (local and/or systemic recurrences), or date of death (Table 1).

All of the tumour samples were formalin-fixed and paraffin-embedded for microscopic examination. The most representative paraffin block of tumour was selected for immunohistochemical analysis. A histological and pathological diagnosis were reviewed by two pathologists (G Alì and G Fontanini), according to the WHO 2004 histological and immunohistochemical criteria (Travis et al, 2004). Disagreements concerning the histological diagnosis were discussed, and after a critical discussion, a mutual agreement was reached. The surgical–pathological staging was performed according to the TNM classification by the International Mesothelioma Interest Group (IMIG) (Rusch, 1995).

Immunohistochemical analyses were performed on 3 μm tissue sections using specific antibodies. Immunoreaction was displayed using the avidin–biotin–peroxidase complex method. Peroxidase activity was visualised with diaminobenzidine. Counterstaining was performed with haematoxylin. The negative controls were carried out by omitting the primary antibodies. Immunostaining was done using a Benchmark immunostainer (Ventana, Tucson, AZ, USA). In all cases, the immunohistochemical evaluation was performed independently by two pathologists (G Alì and G Fontanini) who were blind to the clinicopathological characteristics and treatment of the patients.

For the tryptase and chymase immunohistochemical stainings, sections were incubated with a mouse anti-human tryptase monoclonal antibody (Chemicon International, Temecula, CA, USA), used at a 1 : 1500 dilution, and with a mouse anti-human chymase monoclonal antibody (Chemicon International), used at a 1 : 1600 dilution. Immunostaining for tryptase and chymase was clearly visible as brown deposits within intact MCs and as highly localised extracellular granular material (Figure 1A). Each pathologist identified, under low microscopic power ( × 10 objective lens and × 10 ocular lens), the areas where the MCs were the most intensely accumulated, and then counted positive MCs at × 200 microscopic fields. The average of their counts in three fields was calculated as described earlier (Ibaraki et al, 2005).

For TIL-immunohistochemical staining, sections were incubated with the following antibodies: mouse anti-human Foxp3 monoclonal antibody (clone 236A/E7 diluted 1 : 300; Abcam, Cambridge, UK) (Figure 1B), mouse anti-human CD8 monoclonal antibody (clone C8/144B, ready to use for the Ventana automated slide stainer; Ventana) (Figure 1C), and mouse anti-human CD4 monoclonal antibody (clone 1F6 diluted 1 : 20; Diagnostic BioSystem, Pleasanton, CA, USA). Each pathologist identified, under low microscopic power ( × 25 objective lens and × 10 ocular lens), the areas where the positive cells were most intensely accumulated, and then counted the cells at × 400 microscopic fields. The average of their counts in five fields was calculated as described earlier (Mizukami et al, 2008; Yoshioka et al, 2008).

The microvessel count (MVC) was determined using the anti-CD34 antibody (Ventana Medical System, ready to use for the Ventana automated slide stainer) (Figure 1D). Each pathologist examined the samples and identified the area with the most intense vascularisation (hot spot) under low microscopic power ( × 10 objective lens and × 10 ocular lens). In this area, the number of microvessels were counted and recorded at × 400 ( × 40 objective lens and × 10 ocular lens).

For the VEGF expression, the sections were incubated with an anti-VEGF rabbit polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) used at a 1 : 50 dilution. The expression of VEGF was evaluated as a percentage of positive cells in a total of at least 1000 tumour cells (Ciardiello et al, 2001).

Statistical analysis required the use of the following tests: a Student's t-test for the analysis of the differences between preoperative IL-2-treated patients vs untreated patients; an ANOVA to study the association between risk factors and inflammatory and angiogenetic variables; a Kaplan–Meier test to estimate survival function, and a Cox proportional-hazard model for multivariate survival analysis. The analysis was conducted using R 2.9.1. The level of significance was set at a P-value of less than 0.05.

Ninety-three patients were enroled in this study. On the basis of histological features, the tumours were classified as epithelioid in 46 patients (76.7%), biphasic in 9 patients (15%) and sarcomatoid in 5 patients (8.3%) in the group of patients treated with preoperative IL-2. Preoperative intra-pleural IL-2 was administered to all of the patients without a dose reduction or interruption. All of the patients experienced a fever during treatment and required antipyretic medication (acetaminophen). Out of the 33 mesotheliomas belonging to the group of untreated patients, 25 (75.8%) were epithelioid, 3 (9%) were biphasic and 5 (15.2%) were sarcomatoid. Other clinicopathological characteristics of the series of patients of both groups are summarised in Table 1.

The differences between the two treatment's groups, evaluated by the Student's t-test, are summarised in Table 2.

Regarding the MCs, the IL-2-untreated group showed a significantly lower number of the tryptase MCs in comparison with treated patients' group (P=0.04). No differences between the two groups of patients were observed for the chymase MCs.

Immunohistochemistry revealed an increase in the number of CD8+ lymphocytes in the preoperative-treated group in comparison with the untreated patients (P=0.004). Regarding the CD4+ lymphocytes, the mean value was higher in the group treated with IL-2 than in the untreated group, but the difference was not statistically significant. Evaluation of Foxp3+ lymphocytes revealed a significant higher number in the treated group compared with the untreated group (P=0.02).

As regards MVC, the number of microvessels was significantly lower in the IL-2-treated patients as opposed to untreated patients (P=0.0001). On the contrary, no differences between the two groups of patients were observed for the percentage of VEGF+ neoplastic cells.

Correlations between the analysed parameters and the clinicopathologic characteristics of patients are summarised in Table 3. The MVC was significantly associated with the histologic subtypes of mesothelioma (P=0.0047), with a higher MVC for sarcomatoid and biphasic subtypes than for the epithelioid subtype. No other significant associations were found between the immunological and angiogenetic parameters and the clinicopathological parameters.

In univariate analyses, the histologic subtypes were significantly associated with overall and disease-free survival. The sarcomatoid subtype showed a worse overall survival (P=0.02) and a shorter time-to-progression (P=0.02). No significant association was found between survival and other clinicopathologic parameters (Table 4).

Tryptase-positive MC counts were analysed as a dichotomous variable using the median value to distinguish low (⩽15) from high (>15) MC counts. A statistically significant association was found between the tryptase-positive MC counts and both overall survival (P=0.02) and time-to-progression (P=0.01) (Table 4). A higher tryptase-positive MC count was significantly associated with a better overall survival and a longer time-to-progression (Figure 2A and B).

The Foxp3 count was also analysed as a dichotomous variable using the median value to distinguish tumours with a low or negative Foxp3 count (⩽8.2) and tumours with a high Foxp3 count (>8.2). Univariate analysis showed that the number of Foxp3+ cells was significantly correlated with overall survival (P=0.000002) and time-to-progression (P=0.0039) (Table 4). In fact, high numbers of Foxp3+ lymphocytes were unfavourable to both survival and recurrence (Figure 2C and D).

To assess whether tryptase-positive MC and Foxp3+ lymphocyte counts in the tumours were independent predictors of survival, a Cox proportional-hazard model was used to carry out multivariate analysis. After stratifying for gender, age, histology, performance status and stage, the two variables maintained their independent prognostic roles (Table 5).

The prognostic influence of the combination of tryptase-positive MCs and Foxp3+ lymphocytes was then evaluated. Patients were classified into three groups: I (n=14), Foxp3+ lymphocytes high and tryptase-positive MCs low; II (n=27), both high density or both low density; and III (n=15), Foxp3+ lymphocytes low and tryptase-positive MCs high. Significant differences in overall survival were found among the three groups (P=0.00000017) (Figure 2E). Group III has better overall survival than groups II and I.