Clinical trials of surgical treatment of malignant diseases.

The Dutch Gastric Cancer Study Group Trial was the first clinical phase III trial to be carried out in the field of cancer surgery. In spite of the excellent quality of the trial, it was heavily criticized for the poor quality of the treatment itself. Actually, the hospital mortality after the new surgical treatment (D2 lymph node dissection for gastric cancer) was unacceptably high. In surgical trials, special attention should be paid to quality issues specific to surgery. The first and the most important issue is the quality of treatment given. Reproducibility, homogeneity, and verifiability are the greatest problems in surgical trials. There are also some patient factors. If the patient is old, or fragile, or obese, the results of the surgical treatment can easily be affected by these factors. The surgeon can also be a prognostic factor, especially in complicated procedures or those requiring experience and training. Experience, including postoperative care, and dexterity affect the results. If surgeons do not know how to manage complications, mortality becomes very high. Because blinding is impossible in surgical trials, the treatment may easily be affected by personal preference or prejudice. To minimize the influence of these hampering factors, the procedures should be defined in as detailed a way as possible. If pretrial training or a feasibility study (phase II) is needed, it should be carried out properly for the patients' sake. An excellent design and excellent statistical analysis cannot lead to meaningful results if the quality of treatment is poor. Nonsense in, nonsense out.