Prognosis of the surgical treatment of patients with non-small cell lung cancer (NSCLC) – relation to DNA ploidy

Abstract

Objective: The aim of this study was to evaluate prognostic importance of cell ploidy and proliferation activity in non-small cell lung cancers. Survivals were compared according to the following factors: sex, age, histology, grading, DNA ploidy, tumour size, T factor, N factor and operative procedure. Methods: In a group of 191 patients in whom cytofluorometric examinations had been performed on archival tumour specimens, postoperative recurrences were observed. Results: Postoperative recurrence was observed in 64 (64.6%) of 99 patients with aneuploid tumours and in 35 (38.0%) of 92 with diploid tumours (P<0.001). Overall survival (OS) rates for the group of 92 patients operated for diploid non-small cell lung cancer (NSCLC) at 5 and 10 years were 62 and 51.1%, whereas of other 99, operated for aneuploid tumours 33.3 and 25.9%, respectively (P<0.001). In the former group of patients disease-free survival (DFS) rates at 5 and 10 years were 58.7 and 51.4% but in the latter 29.3 and 26%, respectively (P=0.00014). Significant differences dependent on cell ploidy were also observed in OS and DFS rates of patients operated respectively for SCLC (P=0.0029; P=0.00318) and adenocarcinoma (AC; P=0.0241; P=0.02109). In general, the mean percentage of S-phase cells in non-small cell lung cancers was 14.0% (SD=13.1) in patients who survived 5 years, and 22.4% (SD=15.7) in those who had a recurrence or died (P<0.001). Conclusions: In our opinion the most important finding of our work is that determination of cell ploidy in NSCLC provides a valuable supplement to the TNM stage when evaluating late results of the surgical treatment. However, the paper demonstrates that aneuploidy, although unfavourable, is not an independent prognostic factor in the group of patients with NSCLC and in the subgroups – both with squamous cell carcinoma and adenocarcinoma. Our results show also that the percentage of S-phase cells is an independent, unfavourable prognostic factor in patients treated surgically for non-small cell lung cancer and in the subgroup with squamous cell lung carcinoma.

DNA ploidy, Non-small cell lung cancer, Surgery, Follow-up study

1 Introduction

Patients with non-small cell lung cancer (NSCLC) have the best chance to be cured if radical operative treatment is applied either alone or together with an adjuvant therapy if advanced stages are diagnosed. However, late results of such procedures are not always satisfactory.

A modification introduced in 1997 to the international lung cancer staging system sought to improve correlation among disease stage and treatment outcomes as compared to the previous staging system, but it still does not reliably predict prognosis [1,2].

A possible reason can be the fact that biological properties of a heterogeneous group like non-small cell lung cancers are often different. It seems therefore that to determine biological nature and behaviour of NSCLC would be of utmost importance for the clinic as it would allow to define a group of patients with high risk of recurrence as well as to develop an individual therapeutic program for each patient [3,4].

Although cell ploidy has been shown in a number of studies to be an important independent prognostic factor for many malignant tumours [5,6] including lung cancers [7–10], these relations are still controversial [3,11].

The aim of this study was to evaluate prognostic importance of cell ploidy and proliferation activity in non-small cell lung cancers.

2 Materials and methods

Basing on operation protocols and histopathologic findings regarding primary tumours, removed lymph nodes and operation margins, the lung cancer stages as being determined between 1988 and 1991 according to pTNM classification suggested in 1986 [12] were reclassified retrospectively in patients operated at the Department of Thoracic Surgery of Silesian Medical Academy.

During that time 247 patients were operated for NSCLC (228 men and 19 women). We excluded from consideration a group of 40 patients, whose paraffin-embedded tissue blocks were not suitable for cytofluorometric analysis. Moreover, we excluded 16 (6.5%) surgically treated patients who died in postoperative period (30 days) in a hospital – 12 (8.4%) after pneumonectomy and 4 (3.8%) after lobectomy. Altogether the tumour DNA content was determined in 191 paraffin-embedded tissue specimens from 177 (92.7%) men and 14 (7.3%) women aged 33–74 (av. 54.6) (Table 1) .

Table 1

Patients characteristics

In all patients a complete resection (lobectomy or pneumonectomy) was performed. A complete resection was defined as resection of all macroscopic disease and normal histology of the margin.

The reclassification was performed in accordance with the modification introduced in 1997 by American Joint Committee on Cancer and International Union Against Cancer [13].

Data concerning patients discharged from our Clinic were obtained from Province Thoracosurgical Outpatient Clinic records. Such patients were seen at least four times a year, i.e. every 3 months, within the first 5 years, and every half a year thereafter. Physical examinations, basic laboratory tests, routine X-ray examinations and abdominal USG were performed. If necessary, computed tomography of the chest, abdomen or head was added. Dates of deaths were further verified at Regional Station of Public Records Department for Telecommuni-cation and Computer Science, Ministry of Home Affairs and Administration in Katowice.

We defined local recurrence as a lesion within the same lung or at the bronchial stump; regional recurrence as disease in mediastinal nodes despite mediastinal lymph node dissection during primary operation, lymph node metastasis in contralateral lymph nodes, and/or in supraclavicular regions; single remote metastases (or dissemination) as disease in the distant organs or distant lymph nodes (single or multiple).

2.1 Histopathological analysis

Histological slides were evaluated on the basis of archival material obtained from paraffin-embedded blocks of tumour specimens. Histological type and histological differentiation grade (G) of NSCLC were determined microscopically, and morphological diagnosis was based on the criteria accepted by World Health Organisation (WHO) in 1982 [14]. The degree of histopathological differentiation of the cancers was assessed according to traditional criteria described by Broders [15].

2.2 Preparing for cytofluorometric (FCM) analysis of DNA content

For cytometric examinations of DNA content, suspension of cell nuclei was made from paraffin-embedded tissue blocks, adopting a modified method as described by Hedley et al. [16]. Each analysis comprised at least 10 000 cell nuclei. Ploidy and proliferation were evaluated by means of ModFit program. Classification of DNA histograms in ploidy aspect was based on the criteria recommended by Shankey et al. [17] in their report on the DNA Cytometry Consensus Conference of 1992. Degree of aneuploidy was expressed as DNA index (DI) meaning the ratio of fluorescence intensity for G0/G1 phase cells in a normal population to fluorescence intensity for G0/G1 phase cells in normal population DI=1.0.

Diploid histogram was considered unsuitable for interpretation if its variation coefficient CV (breadth of the fluorescence signal for normal cell population in G0/G1 phase at the middle of its height) exceeded 10%. In case of aneuploid histograms, percentage of S-phase cells was only shown for aneuploid population. Proliferation index was the sum of cells in SG2 and M phases.

2.3 Methods of statistical analysis

Data from case reports were entered into a Microsoft spreadsheet Excel v. 97 to obtain a data base which could be analysed according to standard statistical procedure. In case of normally distributed variables, two mean values were compared using Student's t-test for non-combined variables, whereas three or more groups were compared by multi-mean homogeneity test in single classification. In case of variables showing other than normal distribution, Mann–Whitney U-test was used to compare two groups and Kruskall–Wallis test to compare three or more groups. The probability of overall survival and disease-free survival of patients depending on DNA ploidy was calculated using Kaplan–Meier method (age, tumour size, %SP). Frequency of the variable (sex, histology, grading, T and N factors and pTNM stage) depending on DNA ploidy was compared using chi-square analysis of contingency tests.

Survival curves were determined using Kaplan–Meier method for product limit and with the use of StatSoft program Statistica. F Cox test was used to compare two groups, and chi-square test to compare more than two groups. To see which independent factors had a jointly significant effect on the overall survival rate, we performed Cox’ s multivariate regression analysis.

3 Results

Cytofluorometric analysis of tumour specimens obtained from archival material of 191 patients operated for NSCLC revealed that 99 (51.8%) patients had aneuploid and 92 (48.2%) had diploid tumours.

Mean variation coefficient for aneuploid tumours was 5.87% (range 3.1–10%) and for diploid tumours 6.63% (range 2.4–10%).

We found the relationship between histologic differentiation grades (G) and the increase in NSCLC tumour cell aneuploidy incidence (Table 2) . Of the diploid tumours, only 18 (19.6%) were poorly (G3) differentiated, but of the aneuploid tumours 18 (18.2%) were well (G1) differentiated (P<0.001).

Table 2

Patologic factors and DNA ploidy

Well differentiated cancers were found most rarely among hypertetraploid tumours (5.6%), whereas poorly differentiated cancers in peridiploid and tetraploid groups (14.6% in both).

Aneuploidy incidence and distribution of DNA index for aneuploid tumours in relation to histopathological type of the tumour were not significant (P=0.103).

Moreover there was no relationship between the incidence of tumour aneuploidy, T factor (P=0.338), N factor (P=0.944) and tumour stage (P=0.964), as well as between DNA index in aneuploid tumours and tumour stages (P=0.340).

In general, the mean percentage of S-phase cells in NSCLC was 14.0% (SD=13.1) in patients who survived 5 years, and 22.4% (SD=15.7) in those who had a recurrence or died. The differences were statistically significant (P<0.001).

Median follow-up time was 54 months. Overall survival (OS) rates for the group of 92 patients operated for diploid NSCLC at 5 and 10 years were 62 and 51.1%, whereas of other 99, operated for aneuploid tumours 33.3 and 25.9%, respectively (P<0.001) (Fig. 1a) . In the former group of patients disease-free survival (DFS) rates at 5 and 10 years were 58.7 and 51.4% but in the latter 29.3 and 26%, respectively (P=0.0001). Significant differences dependent on cell ploidy were also observed in OS and DFS rates of patients operated respectively for SCLC (P=0.0029; P=0.0032) (Fig. 1b) and AC (P=0.0241; P=0.0211) (Fig. 1c).

Fig. 1

Influence of cell ploidy on late results of operative treatment in a group of: (a) 191 patients with NSCLC; (b) 126 patients with SCLC; and (c) 50 patients with AC.

Remote metastases (including dissemination) were diagnosed in 46 (71.8%%) of 64 patients operated for aneuploid tumours and in 29 (82.9%) of 35 with diploid NSCLC. Local and/or regional recurrence was found in 18 (18.2%) after aneuploid tumour resection and in six (17.1%) treated for diploid tumours. The most frequently remote metastases were found in the brain (23×), liver (18×) suprarenal gland (17×) and rarely in the second lung (14), and bones (11×). Metastases into other organs (9×) were discovered occasionally.

Cox's proportional hazard ratio revealed that in the group of patients treated surgically for diploid NSCLC grading, N status, T status and %SP are independent, unfavourable prognostic factors for both OS or DFS. Moreover, in the group of patients with aneuploid tumours histologic type showed to be an independent unfavourable prognostic factor when OS or DFS were considered (Table 3) .

Table 3

Multivariate analysis of prognostic factors after surgical therapy of NSCLC patients depending on tumours ploidy

Among patients with aneuploid NSCLC, those with peridiploid (DI 1.1–1.3) and tetraploid tumours (DI 1.9–2.1) had the best prognosis. On the other hand, the shortest survival was observed in the group of patients treated surgically for hypertetraploid (DI>2.1) cancers. The differences had no statistical significance (P=0.142).

In the group of 39 stage I (IA+IB) patients operated for aneuploid NSCLC 56.4% survived 5 years, and in the group of 37 patients treated for diploid tumour – 83.8%. In both groups 45.5 and 75%, respectively survived 10 years (P=0.0056) (Fig. 2a) .

Fig. 2

Influence of tumour cell ploidy on survival after operative treatment in a group of: (a) 76 patients with stage I/IA+IB/NSCLC; (b) 62 patients with stage II/IIA+IIB/NSCLC; and (c) 53 patients with stage IIIA NSCLC.

Analysis of survival rates in stage IA revealed that of those with aneuploid tumours, 71.4% survived 5 years and 42.9% survived 10 years. No patient with a diploid tumour died in this group but all of them (seven) survived 5 and 10 years (P=0.076).

Of 32 stage IB patients treated surgically for aneuploid cancers 51.3% survived 5 years and 46% survived 10 years. In this subgroup, of 30 patients operated for diploid NSCLC, 80% survived 5 years and 69.3% survived 10 years (P=0.019).

Of 29 stage II (IIA+IIB) patients operated for diploid NSCLC 55.2% survived 5 years and 44.3% survived 10 years. Of 33 patients with aneuploid tumours 24.2% survived 5 years and 21% 10 years only (P=0.0169) (Fig. 2b). In the subgroup of stage IIB NSCLC patients the differences were even greater (P=0.008).

In the group of 53 patients in whom stage IIIA was diagnosed, 30.8% survived 5 years of 26 with diploid tumours and 24.6% survived 10 years. However, in the group of 27 patients treated surgically for aneuploid tumour, 7.4% survived 5 years and none survived 10 years (P=0.034) (Fig. 2c).

We did not find any significant differences in survival rates between operated patients with aneuploid NSCLC and those with diploid NSCLC, regardless of histological grade (G) of tumours. Five-years survival rate was found in 97% of patients with diploid and 83.3% with aneuploid G1 tumours (P=0.132). In the group operated for G2 diploid and aneuploid cancers, 5-year survival rate was found respectively, in 48.8 and 27.5% patients (P=0.054). Of the patients with G3 diploid NSCLC, 33.3% survived 5 years but among patients with aneuploid tumours, only 19.5% (P=0.156).

4 Discussion

TNM stage of NSCLC defined on the basis of anatomical principles and follow-up study is a comfortable tool to get understanding of the therapy and allows to select proper kind of the treatment depending on the cancer advancement. However, it is a common observation that this classification is very often insufficient [12,13]. An attempt to seek for more accurate than anatomic only staging of NSCLC is necessary. Knowledge of biologic or molecular characterisation of lung cancer can help in this strategy [2]. It seems that flow cytometric analysis of cellular DNA content and cellular kinetics may provide such information [4,7,10].

However, published papers give conflicting data concerning TNM stage and follow-up study on tumour cell DNA content. Some authors observed shorter survival after aneuploid NSCLC resection, others have shown no effect [3,9,11,18].

In our opinion these conflicting data depended on patients stage evaluation, as well as whether fresh or archival materials were used. Moreover, NSCLC are histologically mixed group of neoplasms. However, we would indicate that all our patients admitted between 1988 and 1991 were consecutive and came from one thoracic center, which minimised selection bias. Clinical, intraoperative and pathological stages were performed by a small and the same group of surgeons and pathologists. Paraffin blocks for each case were chosen from several ones by the pathologist on the basis of histological slides. Thus, possibility of high admixture of non-epithelial cells was minimised. Moreover, all histopathological diagnoses were verified by pathologists from two separate centers, and none of them knew the other diagnosis. If the results were controversial (about 10%) the third pathologist performed the examination once more. His diagnosis was predominant.

We agree with those who consider DNA ploidy as a useful complement of anatomic staging system [8,19,20]. Such approach to the ploidy can be helpful in the follow-up study of surgically treated patients. Moreover, it allows to isolate groups of higher risk patients and to detect cancer recurrence as early as possible.

In aneuploid NSCLC a greater tendency to invade lymphatic vessels and more metastases in regional lymph nodes are noted. Metastases in N1 lymph nodes are reported by some in 18–19% of patients operated for diploid tumours and in 36–54% of patients operated for aneuploid tumours [21,22]. Higher percentage of metastatic mediastinal lymph nodes (59–62%) is also observed in patients with aneuploid tumours, and more often remote metastases are found when compared to patients with diploid tumours [9,10,21]. However, these observations are not always confirmed [7,9].

At the Clinic in the group of patients operated for aneuploid tumours we observed that those with hypertetraploid (DI>2.1) NSCLC as well as with SCLC and AC survived shorter than others. On the other hand, patients with tetraploid AC (DI 1.9–2.1) had better prognosis. However, the differences in survival were not statistically significant. Similar observations were made by others [23].

It is suggested that tetraploidisation of neoplastic cells which had earlier been diploid or peridiploid may lead to aneuploidy together with progressive loss of chromosomal material [24]. This is confirmed by Dalquen's study who detected only four (2.8%) tetraploid (DI 1.9–2.1) and five (3.4%) hypertetraploid tumours (DI 1.7–1.9) among 145 non-small cell lung cancers demonstrating that these tumours are genetically unstable [3]. Moreover, such a low number of diploid or peridiploid tumours observed by him in the study material when compared with aneuploid tumours with relatively low DI (1.4–1.7) resulted in a conclusion that lung cancers are genetically advanced already at the moment of diagnosing, which may account for their aggressiveness and poor late results of treatment [25]. This hypothesis seems to be confirmed by results of the tests performed on archival lung cancer specimens from patients operated at the Clinic: of 191 cases only 11 (5.8%) were tetraploid cancers.

A significant correlation between high DI rate accompanied by higher percentage of S-phase cells and more aggressive biological behaviour of tumours was observed in patients undergoing operation for AC [23].

Investigations performed by other authors on patients with aneuploid tumours also confirmed that changes in higher percentage of S-phase cells were connected with more explicit tendency for metastases and shorter survival [4,7,23]. Furthermore, significantly higher rate of recurrences was observed in patients with tumours showing proliferating fraction higher than 29% in bioptates or exceeding 40% in specimens obtained by open approach [26]. However, these relationships are not confirmed by all authors [9].

5 Conclusions

In our opinion the most important finding of our work is that determination of cell ploidy in NSCLC provides a valuable supplement to the TNM stage when evaluating late results of the surgical treatment. However, the paper demonstrates that aneuploidy, although unfavourable, is not an independent prognostic factor in the group of patients with NSCLC and in the subgroups – both with squamous cell carcinoma and adenocarcinoma. Our results show also that percentage of S-phase cells is an independent, unfavourable prognostic factor in patients treated surgically for non-small cell lung cancer and in the subgroup with squamous cell lung carcinoma.

Dr W. Walker (Edinburgh, Scotland, UK): It is the case, is it not, that some tumours exhibit pleomorphic behaviour with areas of diploidy and areas of aneuploidy. How did you derive your DNA material? Was it from sampling at any particular point in the specimen?

Dr Gawrychowski: It is retrospective material from operated patients in our clinic and archival material from the Department of Pathology of our institution.

Dr P. Macchiarini (Hannover, Germany): The last slide that you showed came from Joe where he made an editorial comment on a very pertinent paper evaluating the different new tools of molecular biology as eventual potential staging tools for lung cancer. So my question is specifically that. You know better than I that you have right now more sensitive markers, like even the mitotic index, or more complex, like the PCNA or the mutant form of the p53, fibroblastic growth factors and so forth. Do you think that these new markers will be getting more sensitive to the staging of these patients, because they were already analysed concerning the disease-free survival. You did not do that analysis; you just took it at outcome, the survival?

Dr Gawrychowski: I think that is the future, but I cannot speak what will be, what factor, because I thought that ploidy will be such, but unfortunately in the Cox analysis it is not confirmed.

Dr M. Tsuboi (Tokyo, Japan): I have two questions. What do you think about the next strategy when you find aneuploidy after surgery?

Dr Gawrychowski: Please repeat.

Dr Tsuboi: How about the next strategy? Is it only surgery and adjuvant treatment? What do you think?

Dr Gawrychowski: In this group?

Dr Tsuboi: Yes.

Dr Gawrychowski: Only surgery.

Dr Tsuboi: Yes, because you find a poor prognostic factor with the aneuploidy. You when you find a poor prognosis suggested in this after surgery, what do you think about it?

Dr Gawrychowski: This work is retrospective. I didn't perform a prospective analysis.

Dr Tsuboi: So when you find that the biopsy specimen shows aneuploidy, how do you treat it in this case?

Dr Gawrychowski: I think that when I have a diploid tumour, we can only cure by surgery, but when aneuploid, it is possible to drug cure, and next, surgery.

Dr Tsuboi: Induction chemotherapy is necessary in that group?

Dr Gawrychowski: Yes, I think.

Dr Tsuboi: Okay.

Dr D. Miller (Atlanta, GA, USA): That was going to be my question. How did you take this from the laboratory into a clinical setting? You're not doing that at the present time, but what we're looking at is, very much like you, we're taking it from an anatomical standpoint, like the T4 men here, but we're taking it to a biological setting of using our tumour markers to do as induction or as adjuvant, and I think that's very important, our main factor, angiogenesis factor as a possibility of treatment, and I was wanting to know, do you-all have plans on doing that? I know this study, this was all tissued from 1988 to 1991, so you've had 11 years to go on to that next step, so I'm just asking, are you planning to do that?

Dr Gawrychowski: I think about prospective analysis.