Cisplatin or carboplatin is commonly used with gemcitabine, docetaxel, paclitaxel, or vinorelbine as chemotherapy doublets in the treatment of advanced non-small-cell lung cancer (NSCLC). Several randomized trials have failed to identify major differences in survival between any of these doublets. This lack of evidence for improvement in survival with any chemotherapy regimen has created a tabula rasa where no more large randomized trials should be conducted without including a genetic analysis. Patients see survival as their major concern, and other considerations, such as cost and quality of life, are relegated to lower positions. Genetic alterations related to the transcription-coupled repair pathway of the nucleotide excision repair system (TC-NER) have revealed the subset of patients who are resistant to cisplatin. TC-NER involves genes that are deficient in rare inborn disorders such as Cockayne syndrome and xeroderma pigmentosum. For a long time, ERCC1 mRNA levels have been known to correlate with DNA repair capacity in various tissues. Levels of DNA cisplatin adducts in peripheral blood and buccal mucosa cells predict chemotherapy response, and high ERCC1 mRNA levels have been related to chemoresistance in ovarian cancer and in malignant lymphocytes from chronic lymphocytic leukemia. Moreover, in some instances, mRNA expression has been correlated with polymorphisms. Overexpression of ERCC1 correlates with poor survival in gemcitabine/cisplatin-treated NSCLC patients. An ongoing customized ERCC1-based chemotherapy trial has been designed based on this knowledge. Patients are randomized to the control arm of cisplatin/docetaxel or to the experimental arm, where docetaxel is combined with cisplatin or gemcitabine according to ERCC1 levels. To date, 86 patients have been included.