| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2006;81:440-447
© 2006 The Society of Thoracic Surgeons
a University of Pittsburgh Medical Center, Presbyterian Shadyside, Heart, Lung and Esophageal Surgery Institute, Pittsburgh, Pennsylvania
b Division of Thoracic and Foregut Surgery, Shadyside Medical Center, Pittsburgh, Pennsylvania
c Cedar Sinai Medical Center, Division of Thoracic Surgery, Los Angeles, California
d Oncotech, Inc, Tustin, California
Accepted for publication August 19, 2005.
* Address correspondence to Dr d'Amato, UPMC Presbyterian Shadyside, Heart, Lung and Esophageal Surgery Institute, Suite C800, 200 Lothrop St, Pittsburgh, PA 15213 (Email: damatota{at}upmc.edu).
Presented at the Forty-first Annual Meeting of The Society of Thoracic Surgeons, Tampa, FL, Jan 24–26, 2005.
| Doctors d'Amato, Landreneau, and McKenna disclose that they have a financial relationship with Oncotech, Inc.
|
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
METHODS: A total of 3,042 NSCLC specimens were cultured in a proliferation assay and tested for resistance to carboplatin, cisplatin, doxorubicin, etoposide, gemcitabine, navelbine, paclitaxel, taxotere, or topotecan. The percentage of cell-growth inhibition measured by 3H-Thymidine uptake was used to determine extreme drug resistance, intermediate drug resistance, or low drug resistance.
RESULTS: Extreme drug resistance or intermediate drug resistance to carboplatin was found in 1,056 of 1,565 NSCLC cultures (68%), to cisplatin in 1,409 of 2,227 (63%), to doxorubicin in 1,101 of 1,471 (75%), to etoposide in 1,581 of 2,505 (63%), to gemcitabine in 594 of 823 (72%), to navelbine in 603 of 1,444 (42%), to paclitaxel in 689 of 1,706 (40%), to taxotere in 273 of 521 (52%), and to topotecan in 280 of 896 (31%).
CONCLUSIONS: Chemotherapy resistance is prevalent among NSCLC clinical cell cultures. This may account for the small survival seen with empiric adjuvant chemotherapy. The use of viable tumor culture for in vitro chemoresistance testing should be considered when formulating a plan of adjuvant therapy for resected NSCLC. Future trials comparing patient survival after tailored versus empiric adjuvant therapy appear justified.
An estimated 172,570 new cases of lung cancer are expected in 2005, accounting for 13% of all cancer diagnoses. Lung cancer is the leading cause of cancer-related death, representing 29% of all cancer-related death [1]. Even though surgical resection is the mainstay of therapy for early-stage nonsmall-cell lung cancer (NSCLC), the expected 5-year survival ranges only from 25% to 65% in patients with completely resected disease [2]. New treatment paradigms focusing on targeted therapy are based on results of laboratory testing. For example, NSCLC tumor epithelial growth factor receptor overexpression and or gene mutation have been utilized to direct therapy with epithelial growth factor receptor small molecule antagonists [3]. In contrast, the empiric use of adjuvant chemotherapy for resected NSCLC, now becoming the standard of care for patients with early stage disease, is being prescribed without prior knowledge of a patient's tumor-specific biology. This "hit or miss" policy has resulted in a marginal survival advantage with empiric adjuvant chemotherapy after resection but results in reliably consistent treatment-related morbidity affecting the majority of patients.
Tumor resistance to chemotherapy may account for the small but significant survival advantage (4% to 15%) observed in patients with resected NSCLC who receive adjuvant chemotherapy [4–6]. As only a minority of NSCLC patients may have an increased survival advantage after adjuvant chemotherapy, and all antineoplastic agents have significant side effects, the majority of patients given adjuvant chemotherapy endure toxicity without added benefit.
Little is known about in vitro tumor resistance to antineoplastic drugs given to patients with resected NSCLC. Tumor chemoresistance testing may be a way to predict clinical unresponsiveness and may help physicians choose alternative therapeutic regimens with potentially greater therapeutic advantage. From a large database of patient-tumor NSCLC cell cultures submitted for in vitro chemoresistance testing (Oncotech, Tustin, California), the prevalence of tumor resistance to commonly used chemotherapeutic agents is reported.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Extreme Drug Resistance Assay Methodology
Accessioned surgical biopsy specimens received in tissue culture media were mechanically and enzymatically disaggregated into single cells and small cellular aggregates. Cell viability was determined by trypan blue dye exclusion. Tissue culture methods were previously described in detail [7]. Briefly, viable cells were suspended in soft-agarose media and plated in duplicate or triplicate at a density of 3 x 104 cells per well in micro titer plates. Single agent antineoplastic agents were added at doses approximating their in vivo peak plasma concentrations [8, 9]. After incubation for 72 hours, 5 µCi of tritiated thymidine (Amersham Biosciences, Piscataway, New Jersey) was added to each well, and cultures were incubated for an additional 48 hours. Cells were lysed with deionized water, agarose-cell suspensions liquefied at 95°C, and contents were harvested onto glass fiber filters. Cellular proliferation was determined by tritiated thymidine–DNA incorporation, measured by liquid scintillation counting and expressed as counts per minute (Fig 1).
|
|
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
The frequency of NSCLC tumors exhibiting either extreme or intermediate chemotherapy resistance is illustrated in Figure 3. Extreme or intermediate drug resistance of cultured human NSCLC tumors exposed to carboplatin was found in 1,056 of 1,565 (68%), to cisplatin in 1,409 of 2,227 (63%), to doxorubicin in 1,101 of 1,471 (75%), to etoposide in 1,581 of 2,505 (63%), to gemcitabine in 594 of 823 (72%), to navelbine in 603 of 1,444 (42%), to paclitaxel in 689 of 1,706 (40%), to taxotere in 273 of 521 (52%), and to topotecan in 280 of 896 (31%) of NSCLC cultures tested.
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Adjuvant Chemotherapy Trials and Toxicity
The efficacy of adjuvant chemotherapy for completely resected NSCLC has been debated for more than a decade. After a meta-analysis suggesting that cisplatin-based chemotherapy without radiation improved the 5-year overall survival rate by 5% and reduced the risk of death by 13% as compared with no adjuvant therapy [10], additional studies were developed and reported. The North American Intergroup (INT 0115) trial compared adjuvant cisplatin plus etoposide and radiation versus adjuvant radiation therapy alone, and showed no benefit from adjuvant chemotherapy [11]. The Adjuvant Lung Project Italy (ALPI) trial showed that patients with completely resected stage I, II, or IIIA NSCLC treated with cisplatin, mitomycin, and vindesine had no statistically significant survival benefit [12]. Toxicity from adjuvant chemotherapy regimens, likely contributed to the lack of benefit noted in these trials. Subsequently, four prospective randomized clinical trials, prompted in part by the results of the meta-analysis, renewed an interest in adjuvant chemotherapy for resected NSCLC.
The International Adjuvant Lung Cancer Trial Collaborative Group evaluated cisplatin-based therapy in 1,867 stage I to III randomized patients who underwent complete resection with a predominance of stage III patients (39%). Patients received a cisplatin doublet (74% receiving 240 mg/m2) and either etoposide (57%), vinorelbine (27%), vinblastine (11%), or vindesine (6%). Seventy percent of patients receiving chemotherapy received an average of 50 Gy of radiation therapy. An absolute 4% increase in survival was noted at 5 years (p < 0.003). Although the International Adjuvant Lung Cancer Trial was not initially intended to have significant power to detect differences in survival for subset analysis, hazard ratios favoring adjuvant chemotherapy versus observation were above the 95% confidence limit, favoring adjuvant chemotherapy only in patients with T3 or T4 tumors, and N2, or stage III, disease [4]. Of the 26% of patients in the chemotherapy arm with incomplete treatment, more than 50% sustained adverse effects. The rate of lethal toxicity from platinum therapy ranged from 2.4% and 0.6% and was not statistically dose dependent.
Recent results presented at ASCO (American Association of Clinical Oncologists) 2004 from the Cancer and Leukemia Group B 9633 trial compared observation only to adjuvant therapy with carboplatin and paclitaxel in 344 randomized stage IB NSCLC patients [5]. The study was limited to T2 N0 patients and radiation therapy was not allowed. A 12% increase in overall survival (p < 0.028) was observed at 4 years with a median follow-up of 34 months. Data on chemotherapy delivery were available on only 124 of 173 (72%). Eighty-five percent of these patients received four doses, yet 35% of this group required dose reductions, only 55% received four cycles at full dose, and adverse event data were available for 149 of 173 patients (86%). This is the only recent North American study to show a survival benefit from adjuvant therapy in completely resected stage IB (T2 N0) disease.
The National Cancer Institute of Canada Clinical Trial Group JBR 10 trial reported by Winton and colleagues [6] at ASCO 2004 further fueled an increased interest in adjuvant chemotherapy for early stage completely resected NSCLC. Patients with stage IB and II (T3 N0 excluded) were stratified based on ras mutation and nodal status. No T1 N0 patients were included and radiation therapy was not allowed. Vinorelbine dosing was reduced owing the high rate of febrile neutropenia. Nineteen percent of patients required hospital admission because of chemotherapy-related toxicity. Only 48% of patients completed four planned cycles of cisplatin-based therapy. An overall improvement in survival of 15% (p < 0.012) was seen in patients receiving adjuvant cisplatin and vinorelbine; however, subset analysis showed only patients with stage II disease in the chemotherapy group had a statistically significant survival advantage.
Further confounding the role of adjuvant chemotherapy for completely resected early stage disease, the Adjuvant Navelbine International Trialist Association (ANITA) trial results recently at ASCO 2005 compared observation only with navelbine plus cisplatin adjuvant therapy in patients with completely resected stage IB (T2 N0), II or IIIA disease. Radiation therapy was permitted. The median percentage of planned does was 56% for navelbine and 76% for cisplatin. Grade 3 or 4 neutropenia occurred with 70% of prescribed doses in 85% of patients. Vinorelabine plus cisplatin significantly improved survival in resected stage II and IIIA disease; however, no benefit was observed in stage IB patients [13].
Adjuvant chemotherapy for completely resected NSCLC, particularly for patients with early stage disease and good performance status, is now an accepted standard of care for resected stage IB, IIA, IIB, and IIIA NSCLC even though only one clinical trial (Cancer and Leukemia Group B 9633) showed improvement in stage IB disease. As noted in these recent adjuvant therapy clinical trials, all antineoplastic drugs exhibited significant toxicity. Although survival advantages were noted, the majority of patients treated did not benefit from adjuvant chemotherapy. Toxic effects resulted in death, often limited dosing, or delayed planned multicycle therapy. Significant toxicity reported in adjuvant chemotherapy clinical trials for NSCLC is summarized in Table 2.
|
Chemoresistance Versus Chemosensitivity
Several clinical correlations support the use of the EDR-Assay versus "chemosensitivity assays" that have been shown to be ineffective to predict a clinical response to chemotherapy. With chemosensitivty assays, the assumption is made that brief exposure to concentrations of chemotherapeutic drugs, at or below clinical peak plasma levels, models the tumors in vivo exposure. However, factors such as interindividual metabolism, tumor vascular supply, and anatomic permeability barriers affect tumor chemotherapy exposure (concentration x time). These potentially important individual patient biologic determinants of response to cytotoxic agents are virtually impossible to assimilate in the in vitro setting. Additionally, pharmacokinetic factors varying from patient to patients such as absorption, activation, and elimination of a cytotoxic drug can cause significant differences in tumor drug exposure, resulting in differences in clinical response even when tumors in an in vitro setting may be equally chemosensitive. These factors significantly limit the accuracy to predict "clinical sensitivity" to any cytotoxic agent [14]. In contrast, the EDR-Assay tumor exposure to antineoplastic agents is long (120 hours) at concentrations that approximate peak plasma levels. Thus, the tumor cell exposure is many times greater than levels achieved in the patient [15, 16]. This results in suprapharmacologic tumor exposure. A specific tumor's growth, measured by tritiated thymidine uptake into DNA, is compared with controls. If the percentage of a specific tumor's growth in the EDR-Assay is one standard deviation below the median inhibition determined for thousands of tumors, then that tumor is considered to be resistant to that drug [14–16]. The EDR-Assay was developed to accurately measure in vitro resistance without reference to either in vitro or clinical sensitivity.
The Bayesian statistical model used to determine the predictive accuracy (posttest probability) of the EDR-Assay is not only a function of tumor exposure, but also the expected response (pretest) probability of a given patient to a specific agent. Kern and Weisenthal [16] in their retrospective chart review of 450 patients correlated assay results and clinical responses. Only one of the 127 patients with extreme drug resistance responded to chemotherapy. In NSCLC, only 2 of 20 patient's tumors exhibiting in vitro intermediate or extreme drug resistance responded to chemotherapy. Subset analysis comparing tumor types expected to be sensitive and those expected to be resistant revealed that the EDR-Assay's ability to identify extreme drug resistance and to predict treatment failure (negative posttest probability of response) was independent of the expected (pretest) probability of response.
Clinical Correlates with Other Solid Tumors
In ovarian cancer, Loizzi and associates [17] showed that patients with platinum-sensitive tumors treated with assay-directed therapy had a highly significant overall survival (90%) compared with patients treated with empiric therapy (70%; p = 0.005), and a highly significant 1-year progression-free survival (assay directed, 68%; empiric treatment, 16%; p = 0.0002). Holloway and colleagues [18] demonstrated that ovarian cancer patients' tumors exhibiting in vitro platinum resistance in the EDR-Assay had a significantly shortened progression-free and overall survival when treated with a platinum-based agent.
For breast cancer patients, Mehta and colleagues [19] showed the median time to disease progression was significantly shorter with extreme or intermediate in vitro drug resistance (48 months) compared with patients with low in vitro resistance (100 months; p = 0.022). Patients whose tumors exhibited in vitro drug resistance had significantly shorter survival in both univariate (p = 0.01) and multivariate analysis (p = 0.04). In vitro drug resistance was associated with an increased relative risk of death (RR = 2.35) similar to that associated with advanced stage or positive lymph node status.
Patients with recurrent malignant glioma were prospectively evaluated for irinotecan resistance [20] in a blinded study to predict the reliability of the EDR-Assay. Median time to progression for IDR/LDR cases was 3 months versus 6 weeks for EDR cases (p = 0.0288, hazards ratio = 3.06). A 13-week median survival for EDR cases was significantly shorter compared with 38 weeks for IDR/LDR cases (p = 0.029). The 100-day survival favored the IDR/LDR cases (p = 0.008).
Current Clinical Application of the EDR-Assay in NSCLC
The prevalence of in vitro chemoresistance in resected NSCLC tumors is shown in this study. Most notable is the frequency of platinum resistance. Therapeutic trends exist wherein a patient's tumor may be resistant to platinum in vitro, yet a platinum-based drug is given as all recent adjuvant trials utilize platinum-based therapy. Although consistent with current treatment algorithms, few patients benefit.
Choosing the second agent in platinum-based therapy by using resistance testing can help physicians "deselect" an agent based on its in vitro ineffectiveness and therefore "potentially" select a more effective therapeutic option. Obviously, this is yet to be validated. However, if a patient's tumor is extremely resistant to etoposide, an alternative second agent such as paclitaxel or navelbine may be substituted with expected similar survival statistics. Both alternatives are accepted second agents. This is well within the "standard of care." Patients periled with metastatic disease or as those undergoing "salvage" therapy may benefit most from chemotherapy deselection when one or more agents may be used, perhaps avoiding toxicity from ineffective second- or third-line therapies. In stage IV lung cancer patients, none of four accepted "empiric" chemotherapy regimens offers a significant survival advantage over another [21].
Limitations of this study include reporting resistance to only single agents when doublet therapy is the clinical standard. Simultaneous exposure of tumor cultures to more than one chemotherapeutic agent at a time was not performed owing to the amount of tissue necessary for analysis such that synergy cannot directly be accessed. Statistical analysis of simultaneous in vitro resistance to more than one agent tested independently is the focus of a separate study in development. Because all tumor histological subtypes were grouped as NSCLC, the prevalence of drug resistance for tumors of different histology was not determined.
Future Directions
A paradigm shift is occurring toward individualized care of patients with resected NSCLC. As the quest to discover and validate targeted therapy continues, chemoresistance testing can bridge the gap between empiric and targeted therapies. A clinical association of clinical cisplatin resistance to high levels of ERCC1 mRNA expression, as recently reported by Rosell and coworkers [22], is an example where identification of in vitro chemotherapy resistance could be applied to identify other genes that impart clinical chemotherapy resistance. Thus, resistance assays should be added to the armamentarium of predictive molecular marker testing, proteomics, and genomic analysis.
Analysis of EDR prevalence in resected NSCLC to first-line chemotherapy agents is the first step toward developing prospective randomized clinical trials of assay directed versus empiric cytotoxic therapy and may help to determine patient-tumor pharmacogenetic profiles for development of targeted therapy in the future.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
DR D'AMATO: The objective here is simply to look at the tumor cells. From an intuitive standpoint, if a tumor cell is resistant to 5 to 20 times the peak serum concentration of a chemotherapeutic agent, why would one expect it to respond in vivo. This leads to the differences really between sensitivities and resistance. This is a resistance assay, which is far more accurate than looking at sensitivity, because, just as you say, there are many host factors that affect sensitivity, whereas this is simply an in vitro resistance assay and is validated only for that.
DR THOMAS A. D'AMICO (Durham, NC): Doctor D'Amato, I admire the direction of your work and I really want to congratulate you on accruing so many samples across three institutions, which, as you know, is not easy to do.
I would like to make two comments. I think it's very important for all the surgeons to realize that this is our future: taking tissue from the OR, testing the biology of the tumor. Involving ourselves in trials like this is very important.
The second comment is, I don't want it to be misunderstood that the results of adjuvant chemotherapy are minor. The 12% and 15% differences in 4-year and 5-year survival in the Cancer and Leukemia Group B 9633 and JBR 10 are the biggest survival advantages ever reported for adjuvant therapy in completely resected solid tumor groups anywhere. So these should be considered huge advances, not small advances. And your point about the size, T1 and T2 versus T3 and T4 in the International Adjuvant Lung Cancer Trial, that study wasn't designed to perform T subgroup analysis. They did that retrospectively. What that analysis shows is that there is no difference in the benefit based on T status, platinum dose, or the platinum doublet. I want to congratulate you on your effort, and we should all be involved in studies like this.
DR MARK J. KRASNA (Baltimore, MD): I enjoyed the presentation. I have just two quick comments. The first one is, it would almost seem from your own presentation that the last conclusion about defining a new protocol with a doublet is unfounded. If anything, what your own data would show thus far is that at least 68%, and as many as 78%, of your specimens were resistant to the agents that we know or we think we know work the most, namely, the platinums. So I think you need to rethink that or perhaps respond why you still think it's a doublet.
Lastly, I guess just to echo a little bit of what Tommy said, I would leave with a note of caution that this idea of a prospective protocol is exactly right, but this schema is a phase III trial. What you really need is a prospective phase II trial to actually justify what you have found in this retrospective analysis, and then if that in fact supports your data, by all means a phase III is the way to go.
DR JOSEPH B. SHRAGER (Philadelphia, PA): I think it's very provocative work. Are there other tumor types in which EDR has been associated with clinical response yet, any other solid tumors?
DR D'AMATO: Yes, there is, most notably in breast carcinoma and ovarian carcinoma recalcitrant to platinum therapy. Highly statistically significant differences were noted in patient survival, particularly in breast and ovarian, and in breast, in multivariate and univariate analysis it appeared to correlate as much as nodal status with marked improvements in survival in assay-directed or in platinum-sensitive patients.
DR SHRAGER: And I guess we're to assume that patient care decisions were never made on the basis of these. I'm amazed that the company would invest in 3,000 specimens or so without having it be used for patient care decisions, meaning that they could have been reimbursed, I guess, for doing it.
DR D'AMATO: I believe there are some rational uses for patient care decisions based on picking agents that are otherwise equally efficacious and choosing one that may not have an extreme drug resistance profile. That stays well within the standard of care. Prospective trials currently underway in ovarian are a blinded trial to platinum resistance, but, to my knowledge, using specific assay-directed chemotherapeutic agents have not been evaluated at this time.
DR SHRAGER: So you could retrospectively look back and maybe give some information about whether—
DR D'AMATO: Yes. That's the topic of a next study, do a retrospective analysis on the tumors that have been received.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. A. d'Amato, R. J. Landreneau, W. Ricketts, W. Huang, R. Parker, E. Mechetner, I.-R. Yu, and J. D. Luketich Chemotherapy resistance and oncogene expression in non-small cell lung cancer J. Thorac. Cardiovasc. Surg., February 1, 2007; 133(2): 352 - 363. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
