Tumor heterogeneity, encompassing genetic, epigenetic, and microenvironmental variables, is extremely complex and presents challenges to cancer diagnosis and therapy. Genomic efforts on genetic intratumor heterogeneity (G-ITH) confirm branched evolution, support the trunk-branch cancer model, and present a seemingly insurmountable obstacle to conquering cancers. G-ITH is conspicuous in clear cell renal cell carcinoma (ccRCC), where its presence complicates identification and validation of biomarkers and thwarts efforts in advancing precision cancer therapeutics. However, long-term clinical benefits on targeted therapy are not uncommon in metastatic ccRCC patients, implicating that there are underlying constraints during ccRCC evolution, which in turn force a nonrandom sequence of parallel gene/pathway/function/phenotype convergence within individual tumors. Accordingly, we proposed a “braided cancer river model” depicting ccRCC evolution, which deduces cancer development based on multiregion tumor genomics of exceptional mTOR inhibitor (mTORi) responders. Furthermore, we employ an outlier case to explore the river model and highlight the importance of “Five NGS Matters: Number, Frequency, Position, Site and Time” in assessing cancer genomics for precision medicine. This mutable cancer river model may capture clinically significant phenotype-convergent events, predict vulnerability/resistance mechanisms, and guide effective therapeutic strategies. Our model originates from studying exceptional responders in ccRCC, which warrants further refinement and future validation concerning its applicability to other cancer types. The goal of this review is employing kidney cancer as an example to illustrate critical issues concerning tumor heterogeneity.