Hematological malignancies continue to pose significant clinical challenges due to their high incidence, heterogeneous biology, and substantial mortality. Although 18F-FDG PET/CT remains the most commonly used molecular imaging modality, its limited specificity can result in false-positive or false-negative findings, especially in indolent or low-metabolism subtypes, thereby hampering accurate diagnosis, staging, and therapeutic evaluation. C-X-C chemokine receptor type 4 (CXCR4) is frequently overexpressed in a broad spectrum of hematologic malignancies and correlates with aggressive disease and unfavorable outcomes. CXCR4-targeted molecular imaging using ^68Ga-pentixafor PET/CT has shown promise for improved disease characterization. This prospective study aims to systematically compare 68Ga-pentixafor PET/CT with 18F-FDG PET/CT in terms of diagnostic performance, staging accuracy, risk stratification, and prognostic relevance in patients with hematological malignancies. Furthermore, the study will incorporate artificial intelligence-based image analysis to enhance lesion detection, automate quantitative assessments, and support personalized clinical decision-making.

Hematological malignancies encompass a heterogeneous group of neoplasms originating from the bone marrow and lymphatic system. Despite advances in therapeutic strategies, the prognosis for many subtypes remains suboptimal, and accurate initial evaluation is critical for appropriate treatment planning. ^18F-FDG PET/CT is routinely used for functional imaging in these patients; however, its diagnostic utility is limited by variable glucose metabolism across disease subtypes and inflammatory uptake, which can lead to misclassification.

CXCR4, a chemokine receptor involved in tumor proliferation, invasion, and microenvironmental interactions, is overexpressed in numerous hematologic malignancies and has emerged as a molecular target for both imaging and therapy. 68Ga-pentixafor, a CXCR4-targeted PET radiotracer, has demonstrated superior lesion detection in preliminary studies, particularly in diseases with low 18F-FDG avidity.

This prospective study will investigate the clinical utility of 68Ga-pentixafor PET/CT in patients with newly diagnosed or relapsed hematological malignancies. Comparative analyses with 18F-FDG PET/CT will be performed to assess concordance, staging accuracy, and association with known risk stratification systems. Longitudinal follow-up will be conducted to evaluate the prognostic relevance of imaging findings.

In parallel, the study will explore the application of artificial intelligence (AI) techniques, including deep learning-based image segmentation and radiomic feature extraction, to enhance diagnostic precision and support risk-adapted management strategies. AI-driven models will be developed and validated to predict disease burden, progression, and survival, thereby contributing to personalized medicine in hematologic oncology.

Inclusion Criteria:

1. Volunteer to participate and sign an informed consent form;
2. 18 ≤ Age ≤ 90 years;
3. Patients with highly suspected, or newly diagnosed, or relapsed hematological malignancies;
4. Willing and able to follow schedule visits, treatment plans and laboratory tests.

Exclusion Criteria:

1. pregnancy or breastfeeding;
2. Allergic to CXCR4-targeted tracers or excipients;
3. Fasting blood glucose level exceeded 11.0 mmol/L prior to injection of 18F-FDG;
4. Those who cannot complete PET/CT scan, including inability to keep supine, claustrophobia, radiation phobia, etc.;
5. Researchers think it is inappropriate to participate in this clinical trial for patients with poor compliance or other unsuitable factors.