Background: Stereotactic dorsal anterior cingulotomy is an established neurosurgical intervention for intractable cancer and non-cancer pain. However, the mechanism by which cingulotomy modulates pain networks for therapeutic effect remains unknown. Methods: This international, multicentre study used stereotactic lesions and patient-derived high angular resolution diffusion imaging (HARDI) and tractography from 26 patients to identify brain connectivity associated with pain alleviation following bilateral anterior cingulotomy. Discovery cohort data (n=14) was used to identify a structural connectivity network disrupted by lesions associated with changes in patient-reported visual analogue scale (VAS) pain scores and morphine equivalent daily dose (MEDD). This network was used to predict ideal lesion locations in an external validation cohort (n=12) using reverse probabilistic tractography. Dominance analyses were performed to identify the contributions of neurotransmitter systems to the topography of our identified pain alleviation network. Results: Chronic pain aetiologies were diverse and were most commonly cancer pain (n=19/26) or neuropathic pain (n=4 peripheral; n=2 central). Pain significantly reduced post-cingulotomy as measured by median decrease in both VAS (50.0% (IQR=70.0-20.0%); P<0.0001) and MEDD (61.1% (IQR=84.3-17.6%); P<0.05). 65.1% patients had a good outcome, defined as ≥30% reduction in VAS. Structural connectivity significantly associated with pain alleviation converged onto a brain network previously implicated in chronic pain maintenance and involving the orbitofrontal cortex, dorsolateral prefrontal cortex, insula, anterior and mediodorsal thalamus, amygdala, striatum, periaqueductal gray, and ventral tegmental area (PFWE<0.05). The Euclidean distance between ideal lesions predicted by connectivity with this network and actual lesions made across the external validation cohort was reliably associated with patient outcome (Spearman’s ρ=-0.43; P<0.01), suggesting that this approach could be used for prospective lesion targeting. Pain alleviation network topography was characterised by opioidergic, histaminergic, and cannabinoidergic neurotransmitter systems (R2adj=0.42), each of which have significant roles in canonical pain processing. Moreover, a functional pain alleviation network derived from 1000 healthy controls using lesion network mapping converged onto the same brain regions. Discussion: We identify the first brain network associated with post-cingulotomy pain alleviation with direct implications for lesion targeting. The overlap of this network with pain-relevant brain regions and neurotransmitter systems provides further insight into the mechanisms underlying chronic pain maintenance. Future neurosurgical interventions for chronic pain should likely aim to modulate this identified network.