SummaryWhen logarithmically growing neuroblastoma cells were exposedin vitrofor 24 hr to a range of concentrations of various antitumour drugs and then cloned in agar to assess colony‐forming ability, the dose‐response curves obtained were of two types only, namely exponential—plateau or exponential. The former category, included hydroxyurea, 1,1′‐propylene bispiperazine‐3,5‐dione (ICRF‐159), vincristine and vindesine, where the plateau was reached before 1 log cell kill, and 4′‐demethylepipodophllotoxin thenylidine glucoside (NSC 122819) (VM‐26) and 4′‐demethyl‐epipodophllotoxin‐9‐(4,6‐oethylidine‐&bgr;‐D‐glucopyranoside (NSC 141540) (VP‐16‐213), where the plateau occurred at approximately 5% survival. Exponential survival curves were obtained with actinomycin D, Adriamycin, bleomycin,cis‐dichlorodiammine platinum (II) (cis‐platinum), dibromodulcitol, 5‐fluorouracil, 4′‐(9‐acridiny‐lamino)methanesulfon‐m‐anisidine (NSC 249992) (mAMSA), melphalan, and peptichemio. A comparison of drug concentrations required to reduce survivalin vitroby 70%, with a 24‐hr drug exposure, was attempted with those concentrations achievable as plasma levels in man after conventional drug therapy. The results obtained, after many and varied assumptions, allowed the division of the drugs tested into three main groups: (1) most effective agents, including mAMSA, VM‐26, and Adriamycin; (2) agents with some activity—vincristine, vindesine, VP‐16‐213,cis‐platinum, melphalan, and peptichemio, although only when the latter three drugs were used at the higher dosage levels; (3) agents with little, if any, activity—actinomycin D, bleomycin, dibromodulcitol, 5‐fluorouracil, hydroxyurea, and ICRF‐159. In terms of attempting clinical correlations, available data suggest definite value in treating neuroblastoma with VM‐26 and Adriamycin [from our group (1)] and also with vincristine, melphalan (high dose), peptichemio, andcis‐platinum [from our group (2)]. No significant clinical activity has been reported for actinomycin D nor for bleomycin in previously treated children, and both these drugs proved negative in this model test system. Among the newer agents tested experimentally, our results suggest that although clinical evaluation of mAMSA, VP‐16‐213, and vindesine might be profitable, dibromodulcitol and ICRF‐159 may prove inactive.SpeculationThe dismal prognosis for the majority of children with neuroblastoma serves to highlight the need for new and more effective therapies. One approach is to attempt to screen experimentallyin vitrofor drugs effective in this disease. Our preliminary studies, using only one cell line derived from one portion of a human tumour, have suggested, with many reservations, some correlations betweenin vitroand clinical data. This work is now being extended to consider other human neuroblastoma cell lines and to monitor retrospectively drug responsesin vitrowith clinical responses. These studies are essential before any attempts are made to extrapolate any of thein vitroresults to humans.