Cerebrospinal fluid (CSF) biomarkers based on the core pathological proteins associated with Alzheimer's disease (AD), i.e., amyloid-β (Aβ) and tau protein, are widely regarded as useful diagnostic biomarkers. However, a lack of biomarkers for monitoring the treatment response and indexing clinical severity has proven to be problematic in drug trials targeting Aβ. Therefore, new biomarkers are needed to track non-Aβ and non-tau pathology. Many proteins involved in the pathophysiological progression of AD have shown promise as new biomarkers. Neurodegeneration- and synapse-related biomarkers in CSF (e.g., neurofilament light polypeptide [NFL], neurogranin, and visinin-like protein 1) and blood (e.g., NFL) aid prediction of AD progress, as well as early diagnosis. Neuroinflammation, lipid dysmetabolism, and impaired protein clearance are considered important components of AD pathophysiology. Inflammation-related proteins in the CSF, such as progranulin, intercellular adhesion molecule 1, and chitinase-3-like protein 1 (YKL-40), are useful for the early detection of AD and can represent clinical severity. Several lipid metabolism-associated biomarkers and protein clearance-linked markers have also been suggested as candidate AD biomarkers. Combinations of subsets of new biomarkers enhance their utility in terms of broadly characterizing AD-associated pathological changes, thereby facilitating precise selection of susceptible patients and comprehensive monitoring of the treatment response. This approach could facilitate the development of effective treatments for AD.