Lowest instrumented vertebra (LIV) selection is critical to preventing complications following posterior spinal arthrodesis (PSA) for thoracolumbar/lumbar adolescent idiopathic scoliosis (TL/L AIS), but evidence guiding LIV selection is limited. This study aimed to investigate the efficacy of PSA using novel unilateral convex segmental pedicle screw instrumentation (UCS) in correcting TL/L AIS, to identify radiographic parameters correlating with distal extension of PSA, and to develop a predictive equation for distal fusion extension using these parameters. We reviewed data (demographic, clinical, radiographic, and SRS-22 questionnaires) preoperatively to 2-years' follow-up for TL/L AIS patients treated by PSA using UCS between 2006 to 2011. 53 patients were included and divided into 2 groups: Group-1 (n=36) patients had PSA between Cobb-to-Cobb levels; Group-2 (n=17) patients required distal fusion extension. A mean curve correction of 80% was achieved. Mean postoperative LIV angle, TL/L apical vertebra translation (AVT), and trunk shift were lower than previous studies. Six preoperative radiographic parameters significantly differed between groups and correlated with distal fusion extension: thoracic curve size, thoracolumbar curve size, LIVA, AVT, lumbar flexibility index, and Cobb angle on lumbar convex bending. Regression analysis optimised an equation (incorporating the first five parameters) which is 81% accurate in predicting Cobb-to-Cobb fusion or distal extension. SRS-22 scores were similar between groups. We conclude that TL/L AIS is effectively treated by PSA using UCS, six radiographic parameters correlate with distal fusion extension, and a predictive equation incorporating these parameters reliably informs LIV selection and the need for fusion extension beyond the caudal Cobb level