Principles of the surgical management of cancer

Carcinogenesis

Neoplasms may be benign or malignant. The cells of benign tumours do not invade surrounding tissues but remain as a local conglomerate. The cells of malignant tumours can directly invade adjacent tissues or enter blood and lymphatic channels, to be deposited at remote sites. This malignant genotype develops as a result of the progressive acquisition of cancer mutations (by point mutation, chromosomal loss or translocation). This progressive accumulation of mutations may lead to the formation of cancer stem cells ( Fig. 8.2 ). These cancer stem cells are pluripotent (i.e., able to give rise to more than one cell type) and produce cells that form the epithelial, structural and vascular components needed for cancer formation. However, cells arising from a cancer stem cell lack the normal response to cell cycle controls and are, therefore, tumour forming. Such cancer stem cells could explain why cancers can relapse or metastasise. The acquisition of the malignant phenotype can be recognised histologically as a tumour develops from a benign adenoma through to a dysplastic lesion and finally into an invasive carcinoma ( Fig. 8.3 ). The concept of tumour progression from benign to malignant phenotype provides the rationale behind screening and early detection programmes. Removing benign or preinvasive lesions will prevent invasive disease.

Invasion and metastasis

Benign tumours rarely threaten life but may cause a variety of cosmetic or functional abnormalities. In contrast, malignant tumours invade and relentlessly replace normal tissues, destroying supporting structures and disturbing function; they can spread to distant tissues (metastasise), eventually causing death. Metastases are cancer deposits similar in cell type to the original cancer found at remote (secondary) sites in the body.

The process of invasion and metastasis is complex ( Fig. 8.4 ) and is dependent on the biology of the tumour. For metastases to occur it would appear that further mutations need to occur in the cancer cells (the metastatic signature). Some tumours metastasise earlier in their clinical course than others. This variation may depend on the tissue of origin of the primary tumour but can also vary widely according to the phenotype of individual tumours. For example, cancer of the breast is thought to metastasise early, and micrometastases are often present but not detectable when the patient first presents. Some patients with apparently localised colorectal cancer are cured by radical surgery, but others receiving the same treatment deteriorate rapidly with metastatic disease.

The mechanisms that control invasion and metastasis are obscure ( Fig. 8.5 ). A variety of enzymes and growth factors are secreted by the tumour cells ( Table 8.2 ); their action facilitates tumour cell invasion and metastasis by degrading extracellular collagens, laminins and proteoglycans.

Clumps of cancer cells can embolise to distant tissues and form metastases. The location for the development of metastases could be a simple mechanical property with organs that have fine capillary beds, such as liver and lung, trapping circulating malignant cells. The survival of metastatic deposits depends on angiogenesis, which is mediated by an imbalance between positive and negative regulatory molecules released by the tumour cells and surrounding normal cells. Cancer cells also secrete prostaglandins, which can induce osteolysis and may promote the development of skeletal deposits.

Natural history and estimate of cure

Calculations based on an exponential model of tumour growth suggest that three-quarters of the lifespan of a tumour is spent in a ‘preclinical’ or occult stage, and that the clinical manifestations of the disease are limited to the final quarter. For cure, every malignant cell must be eradicated and no recurrent tumour should be present during the patient’s lifetime, or evident at death. This rigid definition is rarely attainable, and a normal duration of life without further clinical evidence of disease is generally accepted as evidence of cure, even though microscopic deposits of tumour may still be present.

Measuring and comparing the outcome(s) of cancer treatment can be difficult. Cancer survival data are not normally distributed but skewed, with many events happening early in the study period. Survival data are generally expressed as a time from a predefined starting point (e.g., surgery) to a similarly defined end point (e.g., disease relapse). Other time points may also be used, and so a careful and precise definition of the time period used is essential. In addition, not all patients will have experienced the defined end point by the end of the study period. This phenomenon is known as censoring, and mean survival time will be unknown for a subset of the study group. Other confounding factors such as age and the stage of disease also need to be considered. Hence special methods of data interpretation are required. These various statistical methods of cancer data interpretation and comparison are termed survival analysis.

Measures used in survival analysis include survival and hazard probabilities, Kaplan–Meier equations and graphs ( Fig. 8.6 ), Cox’s proportional hazard models, univariate and multivariate analysis. Survival is the probability that a subject survives from the starting point to the end point of the study period. Hazard is the probability that the subject has a specified event at one particular moment in time. ‘Cure’ rates of individual cancers are assessed by survival rates at various times after treatment. Conventionally, 5- and 10-year intervals are used. Cure rates vary according to the aggressiveness of the disease and the success of treatment. In some patients with cancer (e.g., stomach and lung), metastases grow rapidly and cause death within a few years of clinical presentation. In others (e.g., breast and melanoma), many years may elapse before metastatic spread becomes evident and, even when metastases have occurred, life may be long. It is for this reason that 5-year survival rates cannot provide a satisfactory estimate of cure for all tumours.

Screening

If cancer can be detected before it causes symptoms, then it is generally smaller, has less chance of having metastasised and is therefore more amenable to cure. Detecting benign lesions with malignant potential, preinvasive cancer and invasive malignancy before it becomes symptomatic is called screening ( Fig. 8.7 ). Screening is expensive and its effectiveness in relation to cost must be critically evaluated before routine use ( EBM 8.2 ). Screening is most effective when targeted at specific risk groups and when the screening test has a high level of acceptability to the target population. For successful screening, the test used must be able to detect the cancer at a stage when earlier treatment will lead to fewer deaths from the cancer. In any given population, the likelihood of a cancer being present is generally low (< 1%); hence, the test must be sensitive to detect these relatively rare lesions. The test must also be specific (i.e., have a low false-positive rate); otherwise, individuals will undergo unnecessary investigation or inappropriate treatment. Finally, the proposed treatment of a cancer patient detected by a screening programme must be effective. In the UK, cervical cytology is offered to women on a 3-yearly basis until the age of 60, and mammographic screening ( Fig. 8.8 ) is offered to women between 50 and 64 years on a 3-yearly basis. Other tumour types that might be amenable to screening are listed with their relevant screening tests in Table 8.3 .