Prognostic significance of peritumoural blood and lymphatic vessel invasion by tumour cells in T4 non-small cell lung cancer following induction therapy

Surgical Oncology 1995; 4 : 91-99

Prognostic significance of peritumoural blood and lymphatic vessel invasion by tumour cells in T4 non-small cell lung cancer following induction therapy P. MACCHIARINI, E. DULMET, * V. DE MONTPREVILLE, A . CHAPELlER, J. CERRlNA, F. LE ROY LADURIE AND P. DARTEVELLE Departments of Thoracic and Vascular Surgery and Heart-Lung Transplantation and *Surgical Pathology, H6p ital Marie-Lannelongue (Paris-Sud University), Le Plessis Robinson, France

We investigated the impact of new biological prognostic factors in 28 patients receiving a median of two courses of cisplatin-based chemotherapy with (n = 14) or without (n = 14) radiation and operation for stage IIIB (T4) non-small cell lung cancer (NSCLC). After induction therapy, 5 patients had a complete and 21 a partial response; 2 had a stable disease. A complete resection was made in 26 patients (93%). Five patients (18%) had their primary tumour and involved vestiges completely sterilized. In the remaining 23, the majority of the tumours showed abnormalities in the p53 gene expression (56%), harboured proliferating cells (91%) and induced angiogenesis (91%). Peritumoural blood and lymphatic vessel invasion (PBLVI) by tumour emboli was observed in 6 tumours. With a median follow-up of 25 months, overall 3·year survival was 48%; disease-free survival (DFS) has not been reached yet. The only significant factor influencing DFS in multivariate analysis was PBLVI by tumour cells; PBLVI-positive patients had a significantly higher likelihood ratio (P = 0.000001) of developing metastasis than their PBLVI-negative counterparts. This study documents the prognostic implication of PBLVI by tumour cells in'-4 NSCLC. Surgical Oncology 1994; 4: 91-99. Keywords: non-small cell lung cancer, tumour emboli, peritumoural lymphatic and blood

vessel invasion.

management of patients undergoing such a combined modality treatment. As yet, we have no means of identifying subgroups of patients at high ' risk of developing . metastasis except the presence of few clinical and pathological factors, such as type of response to induction treatment and presence of histological complete sterilization [7J. However, the significance of these factors is not universally agreed upon and they do not explore the biological nature of distant failure. To evaluate ind icator(s) of distant metastasis in patients with 14 NSCLC undergoing induction therapy and operation, we investigated trad itional histological and newer biological parameters recently found to be involved in the early and late events of NSCLC growth and metastasis [8-11 J.

INTRODUCTION To improve the resectability rate and mitigate the risk of developing systemic micrometastasis in locally advanced non -small cell lung cancer (NSCLC), several investigators have evaluated induction chemotherapy with or without radiation before operation [1J. This approach has generated favourable results in stage lilA [2-4J and selected stage 1118 [5, 6J NSCLC patients. Unfortunately, metastatic disease remains the main cause of death and the target issue in the Correspondence: Paolo Macch iarini, MD. Department of Thoracic and Vascular Surgery and Heart-Lung Transplanta tion. Hopltal Marie-Lannelongue (Paris-Sud University). 133. Avenue de la Resistance. 92350 Plessis Robinson. France.

© 1995 Blackwell Science Ltd

91

P. Macchiarini et al.

92 PATIENTS AND METHODS

Patient characteristics Between 1 January 1989 and 30 June 1993, twentyeight consecutive patients with an histologically proven and stage IIIB (T4) NSCLC underwent induction chemotherapy with (n = 14) and without (n = 14) radiation therapy befo re operation; their pre-operative characteristics are listed in Table 1. Twenty-three of them were the subject of a previous publication evaluating the surgical feasibility of this approach [6). Ten tumours originated in the right main stem bronchus and involved the carina (n = 9) or lateral tracheal wall (n = 1). Fifteen were located centrally and invaded the intrapericardial pulmonary artery (n = 6), veins (n = 6) or both (n = 3) and three were apical tumours w ith an extrapulmonary extension into structures of the thoracic inlet. Eight patients (28%) had surgically staged right superior mediastinal (n 6) and left subaortic (n = 2) lvmph : nodes involvement before inclusion into study. Informed consent was obtained from each patient.

=

Pre-operative staging Patients with metastatic pleural effusion, N3 disease and distant metastasis were excluded from the study. Patients were required to have adequate bone marrow, renal and hepatic function and pulmonary reserve. Pretreatment staging for all

Table 1. Patient's pre-operative profile Characteristics Gender Male Female Age (years) Histology Squamous Large cell Adenocarcinoma Side Right Left Previous operation % Predicted post-operative FEV1

Numbers

26

2 54 (range, 38-71) 13

3 12

19 9"

6 (21.4%)

54 .±13.7*

FEV1: forced expiratory volume in the first second. *Mean ± SD of n observations.

patients included history and physical examination, biochemical profile, chest X-ray, computed tomography (CT) of the chest. upper, abdomen and brain , and bronchoscopy. Bone scan was performed in patients with any bone pain. Surgical staging of enlarged mediastinal lymph nodes on chest X-ray or CT scan was made with cervical mediastinoscopy or anterior mediastinotomy. Restaging after induction chemoradiation included chest X-ray, CT scan of the chest and upper abdomen, and bronchoscopy. Except for patients with carinal invasion and diseased med iastinal lymph nodes, mediastinoscopy was not repeated.

Treatment plan Induction chemotherapy consisted of different phase II cisplatin-containing regimens. Eleven patients received two to three cycles of intravenous (i.v.) mitomycin (8 mg m : ", day 1), vindesine (3 mg m - 2 , days 1,8 and 22) and cisplatin (120 mg m- 2 , day 1) (MVP) [12) with (n = 7) or without (n = 4) radiation therapy. Thirteen patients received two cycles of IV fluorouracil (600 mg m - 2 , days 1-4), navelbine (20 mg m - 2 , days 1 and 8), and cisplatin (100 mg m - 2 , day 1) [13) with (n = 3) or without (n == 10) radiation therapy. Four patients received two courses of cisplatin (60 mg m - 2 , days 1-4) and fluorouracil (800 mg m - 2 , days 1-4) and a splitcourse irradiation (10 Gy days 1-4, and 10 Gy, days 8-11) 24-h continuous [14). Except for this last group, radiation therapy was usually delivered at the end of the second treatment course and included 40 Gy targeting the primary lesion, ipsilateral pulmonary hilum, mediastinum, and ipsilateral supraclavicular fossa . Overall, the median delivered dose was 40 Gy. After induction treatment, patients were restaged and those achieving an objective response or stabilization of their disease underwent thoracotomy. A complete response (CR) was defined as the disappearance of all measurable or assessable disease; a partial response (PR), as a > 50% reduction in the sum of the products of the perpendicular diameters of all measured lesions with no evidence of new lesions; stable disease (SD); as a <50% reduction with no evidence of new lesions; and progression (P), as an increase of any measurable or assessable disease. Operation © 1995 Blackwell Science Ltd, Surgical Oncology, 4: 91-99

Vessel invasion by tumour cells was scheduled 4-6 weeks after the completion of chemoradiation and its technical aspects were detailed elsewhere [6]. During operation, resection attempted to remove the entire tumour area present at the time of initial staging and was defined as complete if all known disease was completely resected, the proximal resection margins were microscopically free of tumour, and the highest node area excised or sampled was microscopically not diseased. Incomplete resection was defined as such if gross or microscopically identifiable tumour was left in place. If pathological mediastinal or hilar lymph nodes were present, an extended node dissection was performed. Otherwise, a superior and inferior mediastinal and subcarinal lymph nodes sampling was systematically made. All margins were checked by frozen section [6].

Tumour immunohistochemic studies All surgical specimens were reviewed, without knowledge of patients' outcome, by one pathologist. Positive findings were re-evaluated in a blind fashion by another pathologist for agreement and for consensus on equivocal findings. In the presence of a disagreement, results were recorded as negative. Sections of available lymph nodes were reviewed. Tumour size was assessed on postsurgical specimens. Paraffin-embedded formalin-fixed blocks containing tumour tissue from each patient were analysed. Consecutive 4-Jlm sections were stained by the avidin-biotin-complex (ABC) immunoperoxidase technique [15] using a mouse monoclonal IgG, antibody (Dako Corporation, Santa Barbara, CA, USA) to localize the factor VIIIrelated antigen. The presence of tumour emboli in endothelial-lined channels and their location in the specimen were assessed, and designated as follows : (i) intratumoural lymphatic {ILVI) and/or blood (IBVI) vessel invasion, reflecting the presence of tumour emboli within the primary lesion; (ii) peritumoural lymphatic (PLVI) and/or blood (PBVI) vessel invasion, which encompassed lymphatic and blood vessel invasion peripheral to or at the advancing edge of the primary tumour [8, 9, 16]. Lymphatic vessel invasion (LVI) was distinguished from blood vessel invasion (BVI) based on © 1995 Blackwell Science Ltd, Surgical Oncology, 4: 91-99

previously reported

93

morphologic characteristics

[17, 18). Analysis of p53 human protein was done with the ABC immunoperoxidase techniques using the rabbit polyclonal antibody CM-l (Novocastra, Newcastle, UK) raised against bacterially produced human recombinant p53 [15]. Sections were blocked with normal goat serum for 30 min at room temperature before applying the CM-l antibody 1: 1000 over night. BiotinyJated antirabbit Ig (Vector, Burlingame, CAl application was followed by peroxidase-labelled ABC-complex. Diaminobenzidine was used as chromogen for peroxidase reaction. The presence (positive or negative) and percentage (O-100%) of p53 nuclear immunoreactivity were assessed (11). The areas of tumour containing the most intense angiogenesis (neovascularization) were examined by light microscopy, screened by scanning the tumour sections at low power (40 x and 100 x ) and by identifying the areas with the highest number of discrete microvessels, e.g. capillaries and small venules staining for factor VIII antigen [10]. The areas of high neovascularization were more frequent at the neoplasm margins and graded on a density scale of 0 to 4 +. Microvessels in sclerotic areas within the tumour and adjacent areas of unaffected lung parenchyma were not considered in vessel count and served as internal controls for assessing the quality of staining for factor VIII. After the most vascularize-d area was identified, individual microvessels were counted on 200 x fields and each count was expressed as the highest number of microvessels identified within the area [10]. Cell kinetic studies were determined by estimating the expression of the proliferating cell nuclear antigen (PCNA). PCNA is a highly conserved 36-kD acid nuclear protein which defines and maps the proliferative compartment of proliferative cells in formalin -fixed, paraffinembedded archivial tissue of node-negative NSCLC [19]. Its synthesis increases in late G, phase, reaches its maximum in the S phase, declines during the G2 phase and is absent during the Go and mitotic phases of the cell cycle [20]. Its expression was studied with the monoclonal antibody PC10; this stain ing correlates well with the flow cytometry determination of the Sand G, phases of the cell cycle [21]. Absolute counts of

94

P. Macchiarini et al.

PCNA immunoreactivity were made by scoring a minimum of 5 HPFs and the percentage of proliferating tumour cells was determined (range, 0-100%). Tumours' proliferative activity was classified as low (0-25%), intermediate (26-75%) or high (> 75%) [19].

Five (18%) patients achieved a pathological complete sterilization of their tumours following induction therapy (Table 5). The biological tumoural characteristics of the remaining 23 patients are listed in Table 6. As summarized, the majority of Table 2. Operative findings Characteristics

Statistical analysis Survival was calculated from the date of surgery until death or the date of last follow-up (censored). The disease-free survival (OFS) was considered as the time elapsed from the date of surgery to the first documented evidence of recurrence (local or metastatic). Survival and OFS were estimated by the product-limit method [22), and differences on ·their distribution were evaluated via the log rank test (23), for univariate analysis, and Cox's proportional hazards stepwise model [24), for multivariate analysis. Standard errors associated with the Kaplan-Meier estimator were based on Greenwood's formula (25). Numbers are expressed either as median or as mean ± so of n number of observations. The a priori level of significance was set at a P < 0.05. All tests were two-sided. RESULTS Following induction treatment, 5 patients had a CR, 21 a PR and 2 a SO. Operation was made after a median of 3 weeks (range, 2-5 weeks) from the last day of induction treatment. Operative findings are listed in Table 2; because of the initial location and intraoperative tumoural extension, all but 2 patients requ ired a pneumonectomy. A radical mediastinal node dissection was made in 12 patients. Seventeen (61%) patients had either hilar (n = 11) or mediastinal (n = 6) lymph node involvement at pathological final examination (Table 3). All (93%) but 2 patients with apical lesions were completely resected. Surgical morbidity and mortality is listed in Table 4. Table 5. Comparison between clinical and pathological tumour response following induction therapy

Numbers

Type of operation Tracheal-sleeve pneumonectomy Intrapericardial pneumonectomy Apical resections* Type of involvement Single Multiple Extension Superiorvenacavat Phrenic or vagus nerve Oesophagus+ . Left atrium

10 15

3

5 23

3 9

3 8

*One pneumonectomy and two upper lobectomies. tPolytetrafluoroethylene graft interposition and two plasties were made. +Partial resection of the muscular wall of the oesophaqus. Table 3. Post-surgical nodal stage Type of operation

NO

N1

Tracheal sleeve pneumonectomy Intrapericardial pneumonectomy Apical tumours Total

4 4 2 11

4

6 1 11

N2

Total

2 4 0

10 15 3 28

6

Table 4. Surgical morbidity and mortality No. of patients Morbidity (n = 4, 14%) Atrial fibrillation Haemothorax Pest-pneumonectomy empyema Mortality (n = 4, 14%) Acute distress respiratory syndrome Bronchopleural fistula

2 1 1 2 2

Clinical response

Pathological complete sterilization

Complete response (n = 5) Partial response (n = 21) Stable disease (n = 2) Total

2 3

18

5 (18%)

23 .(82%)

Residual tumour cells

3

2

© 1995 Blackwell Science Ltd, Surgical Oncology, 4: 91-99

Vessel invasion by tumour cells

cells and the difference in their tumour profile as opposed to their PLVI and PBVI negative counterparts is summarized in Table 7. With a median follow-up time of 25 months, the overall 3-year survival, including the 4 postoperative deaths, was 48% (Fig. 1) and the median DFS has not been reached yet (Fig. 2). Postoperative treatment was given to 10 patients and included radiation therapy (median dose, 45 Gy). Six patients died from a local (n = 2) or systemic (n = 4) recurrence first and one, disease-free, from a massive pulmonary embolism (n = 1). Sixteen patients are currently alive and without evidence from disease from 6 to 39 months from operation. The only significant factor influencing survival and DFS in the univariate (Table 8) and multivariate (Fig. 3) analysis was the presence of tumour emboli in both blood and lymphatic peritumoural vessels.

the residual tumours showed abnormalities in the p53 gene expression (56%), harboured proliferating cells (91%) and induced angiogenesis (93%); six patients had both PLVI and PBVI by neoplastic Table 6. Characteristics of the 23 post-surgical specimens with residual tumours. Data are expressed as mean±standard deviation or as medians Values

Characteristics Tumour size (em) Mean, Median p53 expression Present vs. absent Mean (%) Median Tumour proliferative activity* (%) Present vs. absent Mean Median Angiogenesis (present vs. absent) Angiogenesis density (0-4 + scale) Mean Median Angiogenesis count (per 200 x field) Mean Median Peritumoural blood and lymphatic vessel invasion (present vs. absent)

4.6±1.6 4.2 13 vs. 10 37±27.8 30 21 vs. 2 51.4±26.9 50 21 vs. 2

DISCUSSION 1.7±0.7

2

The findings of the present study on patients with stage IIIB (T4) NSCLC treated with induction therapy and operation indicate that blood and lymphatic vessel invasion 'by tumour cells is an independent and significant predictor of DFS irrespective of tumour histology and size, regional

6.6±5.2

6 6 vs. 17

*Determined by proliferating nuclear cell antigen staining.

Characteristics

Histology Squamous Adenocarcinoma Large cells Nodal status

NO N1 N2 Type resection Complete Residual disease (R1)t Tumour size (em) p53 expression (%) Tumour proliferative activity (%)* Angiogenesis density (0-4+ scale) Angiogenesis count (per 200 x field)

95

PLBVI+ (n =6)

PLBVI(n = 17)

2 2 2

Pvalues

9

7

NS

1

2

5

2 2

8 4

NS

5 1 6±1.9 22±24.9 55.8±28.7

16 1 4.1 ± 1.1 30±30.2 49.6±27

0.009 NS NS

1.5±0.5 4.8±3.3

1.53±1.06 6±6

NS NS

*Determined by proliferating nuclear cell antigen staining. tResidual microscopic disease. ©1995 Blackwell Science Ltd, Surgical Oncology, 4: 91-99

NS

Table 7. .Turnour characteristics of the 23 patients with residual 11,011,small cell lung cancer with respect to the presence or absence or peritumourallymphatic and blood vessel invasion (PLBVI) by tumour cells. Data are expressed as mean±standard deviation or as medians

P. Macchiarini et al.

96 1.00

Table 8. Univariate survival and disease-free survival (DFS) analysis (P value) Features

Survival

DFS

Peritumoural lymphatic or blood vessel invasion by tumour emboli (absent vs. present)

0.018

0.0001

0.75

~

:0

0 .0

e

Q.

0.50

0

> .;;

1.00.,----.

:; III

0.25

I I PVI-

0.75

0.00

a

36

24

12

...e

Time (months)

Figure 1. Survival of all 28 patients, including postoperative deaths. Vertical bars indicate patients still alive. 1.00

0.75 ~

:a 0

.

.0

0

~

~ :a .8

0.50

1Il

IJ..

o

0.25 PVI+

• II

"'

,,

0.50

Q.

III

\.L

c

0.25

----T----......24 36

O.OO .. ..------.,.---....... o 12

i

Time (months)

Figure 2. Overall disease-free survival (DFS). Vertical bars indicate patients still alive.

nodal involvement, histological sterilization, intensity and grade of angiogenesis, proliferative activity of residual cells and p53 expression. They also suggest that vessel invasion provides a simple and better tool than other traditional clinical and for improving the pathological parameters prediction of developing metastasis and giving more precise information in making the decision whether adjuvant therapy should be offered to patients whose tumours harbour this feature. Undoubtedly, enthusiasm for our results might be tempered because the sample size is small,

0.00 ........-~--r----1---,.----_r_o 36 12 24 Time (months)

Figure 3. Disease-free survival (DFS) of patients whose tumour harboured peritumoural vessel (lymphatic or blood) invasion (PVI) by neoplastic cells. There was a significant' (P
patients are highly selected and derive from a single institution's experience, and the observed number of relapses limits the statistical power to estimated the type of first recurrence (local versus systemic). However, the fact that the only relapsing patients in this series were those six (21 %) whose neoplasms harboured peritumoural blood and lymphatic vessel invasion by tumour emboli is of paramount importance since it may have several therapeutic implications. First, one might argue that all efforts should be made to detect peritumoural vascular invasion before patients are selected for this multimodality treatment. However, while its detection is relatively easy on tumour specimens at post-surgical evaluation, its discovery on pre-operative tumour biopsies might be difficult since biopsable tumour burdens are usually small. Moreover, the decision to exclude surgery from a chemoradiation ©1995 Blackwell Science Ltd, Surgical Oncology, 4: 91-99

Vessel invasion by tumour cells programme for high-risk patients should be balanced by the fact that the triple modality treatment offered to patients with stage IIIB (T4) NSCLC has a higher therapeutical index than any other approach without induction treatment, inasmuch that we were able to duplicate the 3-year survivals as compared to our retrospective experience without induction treatment [26]. Second, one might consider a more intensive induction regimen before operation to eradicate subclinical metastasis. However, it is our belief that this would further reduce patients' functional status and .increase the sclerofibrotic changes and obliterations 'of hilar and mediastinal planes, enhancing thereby the already heavy surgical morbidity and mortality rates. This comes close to what we would like to accomplish: consideration for a more rigorous post-operative treatment targeting the risk of developing metastasis. We previously demonstrated that the presence of peritumoural blood and lymphatic vessel invasion by NSCLC cells indicates that the neoplasms are in a metastatic phase [8, 9]. Metastasis begins with the local invasion by malignant cells of host stroma within or surrounding the primary tumour. Once tumour cells penetrate a blood vessel or a peripheral lymphatic, they can detach and disseminate; tumour cells that survive in the circulation will then arrest in the microvasculature of the target organ, escape from it and induce angiogenesis in that target organ [27, 28]. In this study of very locally advanced NSCLC, 79% of the tumours were in a pre-metastatic phase. This phase is generally long-lasting, associated with limited tumour growth and a paucity of metastasis, and is common of carcinomas of the cervix [29], bladder [30], breast [31] and peripheral NSCLC [8, 9]. By contrast, the vascular phase begins with the penetration of tumour cells into a blood and/or lymphatic vessel invasion, is short-lasting and associated with rapid tumour growth and increased incidence of metastasis [32, 33]. This distinct figure was observed in our six patients whose T4 NSCLC harboured vessel invasion by tumour cells and this is in line with our earlier experience with T1NOM~ vessel-invasion patients, suggesting that once present, vessel invasion should be targeted by effective post-operative treatment. Interestingly, blood vessel invasion had a more predictive power to detect the likelihood of ©1995 Blackwell Science Ltd, Surgical Oncology, 4: 91-99

97

developing metastasis than another, very promising marker of tumour spread, angiogenesis. It is well known [34] that after a new tumour has attained a size of a few millimetres in diameter (106 tumour cells), further exponential expansion of the tumour-cell population requires the induction of new capillaries blood vessels (angiogenesis) [35], process related to the release of tumour angiogenic factors [36] or deregulation of tumour differentiation [35]. These new capillaries arise from pre-existing capillaries or venules, never from arteries, arterioles, or veins, and represent the consequence of the ingrowth of columns of aligned endothelial cells. Adjacent columns contact to form loops which then develop a lumen, permitting blood flow to begin [33]. The new proliferating capillaries have fragmented basement membranes and their 'leakiness' facilitates their penetration by tumour cells unlike mature vessels [35]. Moreover, rapidly proliferating tumours, like those observed in this report, may force the migration of tumour cells along lines of least resistance [37]. As a consequence, angiogenesis increases the opportunity for single or multiple tumour cells to enter into the circulation [38, 39], particularly above a threshold of neovascularization [11, 40]. Twenty-one (91%) out of 23 lesions with residual tumour cells induced angiogenesis and their median number of neovessels was high compared to other types of NSCLC [8-11]. However, angiogenesis was unable to identify the risk of metastasis, suggesting that, apart from real cell kinetic and extension differences, the presence of vessel invasion by tumour cells is a later marker of the metastatic process of NSCLC and more likely to predict recurrence. Another traditional factor which failed to influence survival and DFS was histological complete sterilization of the specimens. All the patients who had tumour sterilization after induction therapy had major clinical responses and this is in line with recent observations made by Pfister et al. [7]. However, unlike these authors, we did not observe a reduction in recurrence rates in these patients and this is in keeping with early experiences using pre-operative radiation alone [41, 42] but contrasts with more recent series using induction chemotherapy with or without radiation therapy [4, 7, 43]. It may be possible that our data are the results of more detailed evaluation of the biological

98

P. Maeehiarini et a!.

characteristics of the surgical specimens as opposed to all previously reported studies. In conclusion, we have demonstrated that the presence of tumour emboli in the blood and lymphatic peritumoural vessels of patients undergoing induction therapy for stage IIIB (T4) NSCLC is significantly associated with a shorter DFS and a high likelihood of developing metastasis. This provides simple and useful histological information which may aid in formulating treatment policies, especially in the post-operative setting; however, before defining its firm role and routine application, a larger number of patients and longer follow-up time is desirable. Because of the surgical hazards and necessity of large amount of normal and tumoural tissue (affordable only by open biopsy), determination of blood and lymphatic vessel invasion by tumour cells is not recommended before induction therapy.

ACKNOWLEDG EM ENTS The authors thank Professor Michael J . Hardin, PhD, University of Alabama at Birmingham, for statistical assistance.

REFERENCES 1. Van Raemdonk DE, Shneider A, Ginsberg RJ. Surgical treatment for higher stage non-small cell lung cancer. Ann Thorae Surg 1992; 54: 999-1013. 2. Strauss GM, Langer MP, Elias AD, et al. Multimodality treatment of stage IliA non-small-cell lung carcinoma. Clin Oneal 1992; 10: 829-38. 3. Burkes RL, Ginsberg RJ, Shepherd FA, et al. Induction chemotherapy with mitomycin, vindesine, and cisplat in for stage III unresectable non-small-cell lung cancer: results of the Toronto phase II trial . J Clin Oneal 1992; 10: 580-6. 4. Strauss GM, Herndon JE, Sherman DO. Neoadjuvant chemotherapy and radiotherapy followed by surgery in stage lilA non-small-cell lung carcinoma of the lung. J Clin Oneal 1992; 10: 1237-44. 5. Rush VW, Albain KS, Crowley JJ, et al. Surgical resection of stage lilA and stage IIIB non-small-cell lung cancer after induction chemoradiotherapy. J Thorac Cardiovasc Surg1993; 105: 97::"106. 6. Macchiarini P, Chapelier A, Monnet I, et al. Extended operat ions after induction therapy for stage IIIB (T4) non -small cell lung cancer. Ann Thorae Surg 1994; 57: 966-73. 7. Pfister KMW, Kris MG, Gralla RJ, et al. Pathologic

complete response in advanced non-small cell lung cancer following pre-operative chemotherapy: implications for the design of future non-small cell lung cancer combined modal ity trials. J Clin Oneal 1993; 11: 1757-62. 8. Macchiarini P, Fontanini G, Hardin M, et al. Most peripheral, node-negative non-small cell lung cancer have absence of intratumoral and peritumoral blood and lymphatic vessel invasion and low proliferative rates: rationale for treatment with wedge resection alone. J Thorac Cardiavase Surg 1992; 104: 892-9. 9. Macch iarini P, Fontanini G, Hardin, M, et al. Blood vessel invasion by tumour cells predicts recurrence in completely resected Tl NOMO non-small cell lung cancer. J Thorac Cardia vase Surg 1993: 106: 80-9. 10. Macchiarini P, Fontanini G, Hardin, M, et al. Relation of neovascularization to metastasis of non-small cell lung cancer. Lancet 1992; 340: 145-6. 11. Macchiarini P, Fontanini G, Ouimet E, et al. Angiogenesis, an indicator of metastasis in non-small cell lung carcinomas involving the thoracic inlet. Ann Thorec Surg'1994; 57: 1534-9. 12. Martini N, Kris MG, Gralla RJ. The effects of preoperative chemotherapy on the resectability of non-small cell lung carcinoma w ith med iastinal lymph node metastases (N2MO). Ann Thorac Surg 1988; 45: 370-9. 13. Monnet I, Azli N, Voisin S, RufM P, et al. Flourouracil, folinic acid, vinorelbine, cisplatin in non small cell lung cancer: a phase II study. Ann Oneal 1992; 3 (5): 33 (abstract). 14. Rebishung JL, Vannetzel JM, Dartevelle P, Chapelier A. Cyclic concomitant cherno-radiotherapv for primary inoperable non small cell lung cancer. Proc Am Soc Clin Oneal 1991; 10: 254 (abstract) . 15. Hsu SM,' Raine L. The use of avidin -biotin peroxidase complex (ABC) in diagnostic and research pathology. In: DeLessis L, ed. Advances in Immunohistochemistry. New York: Masson 1984: 31-42. 16. Rosen PP. Tumor emboli in intramammary lymphatics in breast cancer: pathological criteria for diagnosis and clinical significance. Pathol Annu 1983; 18: 215-32 (part 2). 17. Weigand RA, Isenberg WM, Russo J, et al. Blood vessel invasion and axiliary lymph node involvement as prognostic indicators for human breast cancer. Cancer 1982; 50: 962-9. 18. Lee AKC, DeLellis RA, Silverman ML, Heatley GI, Wolfe HJ. Prognostic significance of peritumoral lymphatic and blood vessel invasion in node-negative carcinoma of the breast. J Clin Oneal 1990; 8: 1457-65. 19. Fontanini G, Macchiarini P, Pepe S, et al. The expression of proliferating cell nuclear antigen in paraffin sections of peripheral, node -negative nonsmall cell lung cancer. Cancer 1992; 70: 1520-7. 20. Hall PA, Levison DA, Woods AL, et al. Proliferating cell nuclear antigen (peNA) immunolocalization in paraffin sections: an index of cell proliferation with © 1995 Blackwell Science Ltd, Surgical Oncology, 4: 91-99

Vessel invasion by tumour cells

21.

22.

23.

24. 25. 26.

27.

28.

29.

30.

31.

32.

evidence of deregulated expression in some neoplasms. J Patho11990; 162: 285-94. Fontanini G, Pingitore R, Bigini D, et al. Growth fraction in non-small cell lung cancer estimated by proliferating cell nuclear antigen and comparison with Ki-67 labeling and DNA flow cytometry data. Am J Patho/1992; 141: 1285-90. Kaplan El, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958; 53: 457-81. Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SY. Design and analysis of randomized clinical trials requiring prolonged observations of each p~tient: II. Analysis and examples. Br J Cancer 1977; 35: 1-39. Cox DR:"Regression models in life tables. J R Stat Soc (B) 1972;34: 187-220. Kalbfleish JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York: Wiley, 1980. Merleir M, le Brigand H, Rojas A, et al. Etude retrospective de la survie eloiqnee apres exereese d'un cancer bronchique primitif. Chirurgie 1983; 109: 590-7. Fidler IJ, Gersten DM, Hart IR. The biology of cancer invasion and metastasis. Adv Cancer Res 1978; 28: 149-250. Liotta LA, Saidel G, Kleinerman J. The significance of hematogenous tumor cell clumps in the metastatic process. Cancer Res 1976; 36:889-94. Sillman F, Boyce J, Fruchter R. The significance of atypical vessels and necwascularization in cervical neoplasia. Am J Obstet Gynecol1981; 139: 154-9. Chodak GW, Haudenschild C, Gittes RF, Folkman J. Angiogenic activity as a marker of neoplastic and of preneoplastic lesions of the human bladder. Ann Surg 1980; 192: 762-71. Jensen HM, Chen I, deVault MR, lewis AR. Angiogenesis induced by 'normal' human breast tissue: a probable marker for precancer. Science 1982; 218: 293-5. Weidner N, Semple JP, William RW, Folkman J. Tumor angiogenesis and metastasis-correlation in

©1995 Blackwell Science Ltd, Surgical Oncology, 4: 91-99

33.

34.

35. 36.

37.

38.

39.

40.

41.

42.

43.

99

invasive breast carcinoma. N Engl J Med 1991; 324: 1-8. lee AKC, Delellis RA, Silverman Ml, Heatley GI, Wolfe HJ. Prognostic significance of peritumoral lymphatic and blood vessel invasion in node-negative carcinoma of the breast. J Clin Oncol 1990; 8: 1457-65. Folkman J. What is the evidence that tumors are angiogenesis dependent? J Nat! Cancer Inst 1990; 82: 4-6. Hart IR, Saini A. Biology of tumour metastasis. Lancet 1992; 339: 1453-7. Folkman J, Watson K, Ingberg D, Hanahan D. Induction of angiogenesis during the transition from hyperplasia to neoplasia. Nature ·1989; 339: 58-61. Eaves G. The invasive growth of malignant tumors as a purely mechanical process. J Pathol 1972; 109: 233-7. Liotta LA, Steeg PS, Stetler-Stevenson WG. Cancer metastasis and angiogenesis: an imbalance of positive and negative regulation. Cell 1991; 64: 327-36. Liotta t, Kleinerrnan J, Saidel G. Quantitative relationships of intravascular tumour cells, tumour vessels, and pulmonary metastasis following tumor implantation. Cancer Res 1974; 34: 997-1004. Horak ER, leek R,Klenk N, et al. Angiogenesis, assessed by platelet/endothelial cell adhesion molecule antibodies, as indicator of node metastases and survival in breast cancer. Lancet 1992; 340: 1120-224. Bromley LL, Szur L. Combined radiotherapy and resection for carcinoma of the bronchus. Lancet 1955; 937-41. Shields TW, Higgins GA, lawton R, et al. Preoperative x-ray therapy as an adjuvant in the treatment of bronchogenic carcinoma. J Thorac Cardiovasc Surg 1970; 59: 49-61. Faber lD, Kittle CF, Warren WF, et al. Preoperative chemotherapy and irradiation for stage III non-small cell lung cancer. Ann Thorac Surg 1989; 47: 669-677.