Dormancy and surgery-driven escape from dormancy help explain some clinical features of breast cancer.

To explain bimodal relapse patterns observed in breast cancer data, we have proposed that metastatic breast cancer growth commonly includes periods of temporary dormancy at both the single cell phase and the avascular micrometastasis phase. The half-lives of these states are 1 and 2 years respectively. We also suggested that surgery to remove the primary tumor often terminates dormancy resulting in accelerated relapses. These iatrogenic events are very common in that over half of all metastatic relapses progress in that manner. Assuming this is true, there should be ample and clear evidence in clinical data. We review here the breast cancer paradigm from early detection, through treatment and follow-up, and consider how dormancy and surgery-driven escape from dormancy would be observed. We examine mammography data, effectiveness of adjuvant chemotherapy, heterogeneity and aggressiveness, timing of surgery within the menstrual cycle and racial differences in outcome. Dormancy can be identified in these diverse data but most conspicuous is the sudden escape from dormancy following primary surgery. These quantitative findings provide linkage between experimental studies of tumor dormancy and clinical efforts to improve patient outcome.

The effects of surgery on tumor growth: a century of investigations.

A few clinical investigations suggest that while primary breast cancer surgical removal favorably modifies the natural history for some patients, it may also hasten the metastatic development for others. The concepts underlying this disease paradigm, i.e. tumor homeostasis, tumor dormancy and surgery-driven enhancement of metastasis development, have a long history that is reviewed. The review reveals the context in which these concepts were conceived and structured to explain experimental data and shows that they are not so new and far fetched. The idea that surgical cancer resection has both beneficial and adverse effects upon cancer spread and growth that result from the modulation of tumor dormancy by the resection should be considered a potentially fruitful working hypothesis.

Breast cancer recurrence dynamics following adjuvant CMF is consistent with tumor dormancy and mastectomy-driven acceleration of the metastatic process.

PURPOSE: The aim of this study was to better understand human breast cancer biology by studying how the timing of metastasis following primary resection is affected by adjuvant CMF (cyclophoshamide, methotrexate, 5-fluorouracil) chemotherapy. PATIENTS AND METHODS: Discrete hazards of recurrence and recurrence risk reductions for treated patients relative to controls were analyzed for all patients enrolled in two separate randomized clinical trials [study 1 (386 women): no further treatment versus 12 cycles of CMF; study 2 (459 women): six versus 12 cycles of CMF] and a historical group (396 women: surgery alone) of axillary node-positive patients undergoing mastectomy. RESULTS: (i) Nearly all CMF benefit occurs during the first 4 years following resection/chemotherapy. (ii) The CMF recurrence rate reduction is largely restricted to two specific spans. These temporally separate recurrence clusters occur during the first and third year of follow-up, while the second-year recurrences are weakly affected. (iii) Prolonging adjuvant treatment from 6 to 12 months partially alters this recurrence timing, without appreciably affecting the overall recurrence rate. (iv) These effects upon the dynamics of post-resection occurrence are menopausal status-independent. CONCLUSIONS: At least two different therapeutically vulnerable proliferative events, resulting in clinical appearance of two metastasis temporally distinct clusters of post-resection cancer recurrence, apparently occur during the administration of adjuvant chemotherapy. Metastases that transpire outside of these temporal windows are refractory to adjuvant therapy. The dynamics of both post-treatment recurrence risk and CMF effectiveness are similar for both pre- and postmenopausal women, suggesting that post-resection mechanisms by which chemotherapy prevents metastases are similar, but of different magnitude in pre- and postmenopausal women. These findings are consistent with a metastasis model that includes tumor dormancy in specific micrometastatic phases (single cells and avascular foci) and with the acceleration of the metastatic process by the surgical resection of the primary breast cancer.

Computer simulation of a breast cancer metastasis model.

Recent analysis of relapse data from 1173 untreated early stage breast cancer patients with 16-20 year follow-up shows that frequency of relapse has a double peaked distribution. There is a sharp peak at 18 months, a nadir at 50 months and a broad peak at 60 months. Patients with larger tumors more frequently relapse in the first peak while those with smaller tumors relapse equally in both peaks. No existing theory of tumor growth predicts this effect. To help understand this phenomenon, a model of metastatic growth has been proposed consisting of three distinct phases: a single cell, an avascular growth, and a vascularized lesion. Computer simulation of this model shows that the second relapse peak can be explained by a steady stochastic progression from one phase to the next phase. However, to account for the first relapse peak, a sudden perturbation of the development at the time of surgery is necessary. Model simulations predict that patients who relapse in the second peak would have micrometastases in states of relatively low chemosensitivity when adjuvant therapy is normally administered. The simulation predicts that 15% of T1, 39% of T2, and 51% of T3 staged patients benefit from adjuvant chemotherapy, partially offsetting the advantage of early detection. This suggests that early detection and adjuvant chemotherapy may not be symbiotic strategies. New therapies are needed to benefit patients who would relapse in the second peak.

Proposal for a new model of breast cancer metastatic development.

BACKGROUND: The commonly accepted theory of breast cancer metastatic development assumes continuous tumor growth from tumor seeding until documentation of clinical recurrence. In particular, Gompertzian growth kinetics is currently the theoretical cornerstone of the natural history of breast cancer, and has been widely utilized for planning treatments. MATERIALS AND METHODS: To verify agreement between findings and the implications of the continuous growth model, several published papers about the natural history of breast cancer after removal of the primary tumor were reviewed. Also, findings from animal models concerning metastasis biology were considered. RESULTS: The continuous growth model failed in important ways upon this critical reappraisal. As an alternative, the tumor dormancy hypothesis was considered to provide a more reasonable description of tumor recurrence. Moreover, primary tumor removal was revealed as a potentially perturbing factor for metastasis development. CONCLUSIONS: A new general outline of metastatic development of breast cancer incorporating tumor dormancy in specific micrometastatic phases, stochastic transitions between them, and start signals from surgery for micrometastatic growth was designed. The proposed model suggests new views concerning scheduling of current chemotherapy, new treatment approaches aimed at keeping micrometastases in a dormant state for the patient's entire life, and the careful reappraisal of the timing of surgery within the multimodal treatment of operable breast cancer.

Dose response in the treatment of breast cancer.

We present evidence that drugs that are effective in causing regression of metastatic breast cancer in a dose-dependent fashion also cause a dose-dependent reduction of cytokines produced by the immune system. This may be an important factor in the occasional induction of a complete remission.

Computer model challenges breast cancer treatment strategy.

The breast cancer treatment failure rate remains unacceptably high. The current breast cancer treatment paradigm, based primarily on Gompertzian kinetics and animal models, advocates short-course, intensive chemotherapy subsequent to tumor debulking, citing drug resistance and host toxicity as the primary reasons for treatment failure. To better understand treatment failure, we have studied breast cancer from the perspective of computer modeling. Our results demonstrate breast cancers grow in an irregular fashion; this differs from the Gompertzian mode of animal models and thus challenges the validity of the current paradigm. Clinical and laboratory data support the concept of irregular growth rather than the common claim that human tumors grow in a Gompertzian fashion. Treatment failure mechanisms for breast cancer appear to differ from those for animal models, and thus treatments optimize on animal models may not be optimal for breast cancer. A failure mechanism consistent with our results involves temporarily dormant tumor cells in anatomical or pharmacological sanctuary, which eventually result in aggressive metastatic disease.

An alternative approach for treatment of breast cancer.

Since adjuvant chemotherapy and hormonal therapy generally extend disease free survival in breast cancer rather than provide a cure, we have examined the current breast cancer paradigm. Heterogeneity is a fundamental characteristic of breast cancer tissue and a well recognized aspect of the disease. There are variations in natural history, histopathology, biochemistry and endocrinology, and molecular biology of cancer tissues and cells within the tissues. A variety of data indicate that growth kinetics are also variable, not only from tumor to tumor, but also during the natural history of an individual's tumor. To better understand kinetic heterogeneity, a stochastic numeric computer model of the natural history of breast cancer has been developed. To be consistent with inter- and intratumor kinetic heterogeneity and with late relapse, the model predicts that tumors grow in an irregular fashion with alternating periods of growth and periods of dormancy rather than the generally accepted modified exponential, or Gompertzian fasion. The prediction of irregular growth has been compared to data relevant to growth characteristics of human breast cancer. Much data support the concept of irregular kinetics and temporary dormancy rather than steady, Gompertzian growth of human breast cancer. Thus, in addition to drug resistance, kinetic heterogeneity may help explain the limited impact that traditional chemotherpeutic treatment has had on mortality from breast cancer. Although the mechanisms underlying irregular growth need to be better understood, non-Gompertzian growth kinetics indicates that there may be alternative approaches for breast cancer treatment.

Is Gompertzian or exponential kinetics a valid description of individual human cancer growth?

It is generally accepted that human cancers grow in an exponential or Gompertzian manner. This assumption is based on analysis of the growth of transplantable animal tumors and on averages of tumor growth in human populations. A computer model of breast cancer in individual patients has raised some doubts about this assumption. The computer model predicts an irregular pattern of tumor growth that incorporates plateaus or dormant periods separated by Gompertzian growth spurts. Since growth patterns involving plateaus are not predicted by conventionally accepted exponential or Gompertzian kinetics, sufficient documentation of their existence may be regarded as some evidence that the computer model is correct. The literature has been surveyed to identify growth patterns specifically predicted by the model. The literature contains clinical evidence from individual patients of this growth pattern in primary breast, large intestine and rectum, and pulmonary cancers and metastatic pulmonary cancer. Much data, including the only breast data, are not consistent with exponential or Gompertzian kinetics but are explainable by irregular growth kinetics. Exponential growth is valid for some tumors and for short times, but there are many papers citing significant deviations from that growth. Exponential growth may accurately describe averages of human tumor growth and growth of multipassaged experimental tumors, but it is not valid for all individual tumors.

Prospective computerized simulation of breast cancer: comparison of computer predictions with nine sets of biological and clinical data.

A computer program which accepts clinically relevant information can be used to predict breast cancer growth, response to chemotherapy, and disease-free survival. The computer output is patient individualized because the program is highly iterative and simulates up to 2500 patients with exactly the same clinical presentation. Computer predictions have been compared to a broad spectrum of breast cancer data, and a high degree of correlation has been established. There are numerous significant clinical implications which can be derived from the computer model. Among these are the following. (a) Breast cancer tumors do not grow continuously but may have up to five growth plateaus each lasting from a small fraction of a year up to approximately 8 yr. (b) Adjuvant chemotherapy, such as 6-mo treatment with cyclophosphamide-methotrexate-5-fluorouracil, does not eradicate tumors but just reduces the number of viable cells by a factor of 10 to 100 and sets the eventual growth back by several years. This may partially explain why the age-adjusted death rate from breast cancer has not changed in the past 50 yr. (c) The computer model challenges the underlying principles in support of short-term intensive adjuvant chemotherapy, namely Gompertzian kinetics and genetically acquired tumor resistance to drugs. (d) The computer model questions the evidence opposing long-term maintenance chemotherapy protocols and suggests that maintenance protocols should be reexamined.

A stochastic numerical model of breast cancer growth that simulates clinical data.

A new stochastic numerical model of breast cancer growth is developed. First, the model suggests that Gompertzian kinetics does apply but that from time to time, in random fashion, there occurs a spontaneous change in the growth rate or rate of decay of growth, such that the overall growth pattern occurs in a stepwise fashion. According to the model, the average time for the tumor burden to increase from one cell to detection is probably in the range of 8 years. Secondly, the model suggests that there is a linear relationship between the number of axillary lymph nodes positive for metastasis at diagnosis and the number of other metastatic sites. This can be described mathematically by the equation S = 0.24 + 0.35N where S is the number of other metastatic sites and N is the number of positive lymph nodes. The model has been verified by simulating three data sets: (a) the survival times of untreated breast cancer patients as described by Bloom et al. [Br. Med. J., 2: 213-221, 1962]; (b) the growth rates of breast cancers immediately prior to diagnosis as described by Heuser and Spratt [Cancer (Phila.), 43: 1888-1894, 1979]; and (c) the disease-free survival time postmastectomy as described by Fisher et al. [Surg. Gynecol. Obstet., 140: 528-534, 1975]. This model could have implications concerning the overall treatment rationale for breast cancer.