Self-inspired learning for denoising live-cell super-resolution microscopy.

Every collected photon is precious in live-cell super-resolution (SR) microscopy. Here, we describe a data-efficient, deep learning-based denoising solution to improve diverse SR imaging modalities. The method, SN2N, is a Self-inspired Noise2Noise module with self-supervised data generation and self-constrained learning process. SN2N is fully competitive with supervised learning methods and circumvents the need for large training set and clean ground truth, requiring only a single noisy frame for training. We show that SN2N improves photon efficiency by one-to-two orders of magnitude and is compatible with multiple imaging modalities for volumetric, multicolor, time-lapse SR microscopy. We further integrated SN2N into different SR reconstruction algorithms to effectively mitigate image artifacts. We anticipate SN2N will enable improved live-SR imaging and inspire further advances.

A compressive seeding algorithm in conjunction with reordering-based compression.

MOTIVATION: Seeding is a rate-limiting stage in sequence alignment for next-generation sequencing reads. The existing optimization algorithms typically utilize hardware and machine-learning techniques to accelerate seeding. However, an efficient solution provided by professional next-generation sequencing compressors has been largely overlooked by far. In addition to achieving remarkable compression ratios by reordering reads, these compressors provide valuable insights for downstream alignment that reveal the repetitive computations accounting for more than 50% of seeding procedure in commonly used short read aligner BWA-MEM at typical sequencing coverage. Nevertheless, the exploited redundancy information is not fully realized or utilized. RESULTS: In this study, we present a compressive seeding algorithm, named CompSeed, to fill the gap. CompSeed, in collaboration with the existing reordering-based compression tools, finishes the BWA-MEM seeding process in about half the time by caching all intermediate seeding results in compact trie structures to directly answer repetitive inquiries that frequently cause random memory accesses. Furthermore, CompSeed demonstrates better performance as sequencing coverage increases, as it focuses solely on the small informative portion of sequencing reads after compression. The innovative strategy highlights the promising potential of integrating sequence compression and alignment to tackle the ever-growing volume of sequencing data. AVAILABILITY AND IMPLEMENTATION: CompSeed is available at https://github.com/i-xiaohu/CompSeed.

Mapping secretome-mediated interaction between paired neuron-macrophage single cells.

Neuron-immune interaction through secreted factors contributes significantly to the complex microenvironment in the central nervous system that could alter cell functionalities and fates in both physiological and pathological conditions, which remains poorly characterized at the single-cell level. Herein, using a spatially patterned antibody barcode microchip, we realized the mapping of 12 different secretomes, covering cytokines, neurotrophic factors (NFs), and neuron-derived exosomes (NDEs) from high-throughput, paired single cells (≥ 600) simultaneously under normal conditions and an Alzheimer's disease (AD) model induced with amyloid beta protein 1-42 (Aß1-42). We applied the platform to analyze the secretion profiles from paired neuron-macrophage and neuron-microglia single cells with human cell lines. We found that pairwise neuron-macrophage interaction would trigger immune responses and attenuate neuron cells' secretion, while neuron-microglia interaction generally results in opposite outcomes in secretion. When neuron cells are induced with Aß1-42 protein into the AD model, both neuron-macrophage and neuron-microglia interactions lead to increased cytokines and NDEs and decreased NFs. Further analysis of AD patients' serum showed that NDEs were significantly higher in patients' samples than in the control group, validating our observation from the interaction assay. Furthermore, we resolved previously undifferentiated heterogeneity underlying the secretions from single-neuron cells. We found that the NDE and NF secretion was less dependent on the paracrine signaling between one another and that secretions from neuron cells would attenuate after differentiation with Aß1-42. This study demonstrates the mapping of the different secretomes from paired neuron-immune single cells, providing avenues for understanding how neurons and immune cells interact through the complex secretome network.

Automated Droplet Ejection from a Digital Microfluidics Sample Pretreatment Device Enables Batch-Mode Chemiluminescence Immunoassay.

Digital microfluidics (DMF) features programmed manipulation of fluids in multiple steps, making it a valuable tool for sample pretreatment. However, the integration of sample pretreatment with its downstream reaction and detection requires transferring droplets from the DMF device to the outside world. To address this issue, the present study developed a modified DMF device that allows automated droplet ejection out of the chip, facilitated by a tailor-designed interface. A double-layered DMF microchip with an oil-filled medium was flipped over, with a liquid infusion port and a liquid expulsion port accommodated on the top working PCB plate and the bottom grounded ITO plate, respectively, to facilitate chip-to-world delivery of droplets. Using chemiluminescent immunoassay (CLIA) as an illustrative application, the sample pretreatment was programmed on the DMF device, and CLIA droplets were ejected from the chip for signal reading. In our workflow, CLIA droplets can be ejected from the DMF device through the chip-to-world interface, freeing up otherwise occupied electrodes for more sample pretreatment and enabling streamlined droplet microreactions and batch-mode operation for bioanalysis. Integrated with these interfacing portals, the DMF system achieved a single-channel throughput of 17 samples per hour, which can be further upscaled for more productive applications by parallelizing the DMF modules. The results of this study demonstrate that the droplet ejection function that is innovated in a DMF sample pretreatment microsystem can significantly improve analytical throughput, providing an approach to establishing an automated but decentralized biochemical sample preparation workstation for large-scale and continuous bioanalysis.

Hydroxyl-functionalized pillar[5]arene with high separation performance for gas chromatography.

Here we report the first example of employing hydroxyl-functionalized pillar[5]arene (P5A-C10-OH) as stationary phase for capillary gas chromatographic (GC) separations. The statically coated P5A-C10-OH capillary column possessed moderate polarity and column efficiency of 3233 plates per m determined by n-dodecane. As a result, the P5A-C10-OH column exhibited high-resolution capability for the mixture of 17 analytes from apolar to polar nature. Importantly, it exhibited advantageous performance for high resolution of the challenging isomers of bromonitrobenzene, chloroaniline, bromoaniline, iodoaniline and dimethylaniline with good peak shapes over the P5A-C10 and commercial HP-35 columns. In addition, eight cis-/trans-isomers with diverse types were baseline separated on the P5A-C10-OH column. And the application of detecting isomeric impurities in real samples gave strong evidence of its potential and feasibility for the viable GC analysis.

Ester-functionalized pillar[6]arene as the gas chromatographic stationary phase with high-resolution performance towards the challenging isomers of xylenes, diethylbenzenes, and ethyltoluenes.

This work presents the first example of the utilization of polar ester group functionalized pillar[6]arene (P6A-C10-OAc) as a stationary phase for capillary gas chromatographic (GC) separations. The statically coated P6A-C10-OAc column showed a high column efficiency of 5393 plates/m and moderate polar nature. Its resolving capability and retention behaviors were investigated for a mixture of 20 analytes and more than a dozen isomers from apolar to polar in nature. As evidenced, the P6A-C10-OAc column achieved high-resolution separations of all the analytes and good inertness. Importantly, it exhibited distinctly advantageous performance for high resolution of the challenging isomers of xylenes, diethylbenzenes, ethyltoluenes, and halobenzenes over the commercial HP-5 (5% phenyl dimethyl polysiloxane), HP-35 (25% phenyl dimethyl polysiloxane), and PEG-20M (polyethylene glycol) columns.

Synergistic Engineering of CoO/MnO Heterostructures Integrated with Nitrogen-Doped Carbon Nanofibers for Lithium-Ion Batteries.

The integration of heterostructures within electrode materials is pivotal for enhancing electron and Li-ion diffusion kinetics. In this study, we synthesized CoO/MnO heterostructures to enhance the electrochemical performance of MnO using a straightforward electrostatic spinning technique followed by a meticulously controlled carbonization process, which results in embedding heterostructured CoO/MnO nanoparticles within porous nitrogen-doped carbon nanofibers (CoO/MnO/NC). As confirmed by density functional theory calculations and experimental results, CoO/MnO heterostructures play a significant role in promoting Li+ ion and charge transfer, improving electronic conductivity, and reducing the adsorption energy. The accelerated electron and Li-ion diffusion kinetics, coupled with the porous nitrogen-doped carbon nanofiber structure, contribute to the exceptional electrochemical performance of the CoO/MnO/NC electrode. Specifically, the as-prepared CoO/MnO/NC exhibits a high reversible specific capacity of 936 mA h g-1 at 0.1 A g-1 after 200 cycles and an excellent high-rate capacity of 560 mA h g-1 at 5 A g-1, positioning it as a competitive anode material for lithium-ion batteries. This study underscores the critical role of electronic and Li-ion regulation facilitated by heterostructures, offering a promising pathway for designing transition metal oxide-based anode materials with high performances for lithium-ion batteries.

Evaluation of DDA Library-Free Strategies for Phosphoproteomics and Ubiquitinomics Data-Independent Acquisition Data.

Phosphoproteomics and ubiquitinomics data-independent acquisition (DIA) mass spectrometry (MS) data is typically analyzed by using a data-dependent acquisition (DDA) spectral library. The performance of various library-free strategies for analyzing phosphoproteomics and ubiquitinomics DIA MS data has not been evaluated. In this study, we systematically compare four commonly used DDA library-free approaches including Spectronaut's directDIA, DIA-Umpire, DIA-MSFragger, and in silico-predicted library for analysis of phosphoproteomics SWATH, DIA, and diaPASEF data as well as ubiquitinomics diaPASEF data. Spectronaut's directDIA shows the highest sensitivity for phosphopeptide detection not only in synthetic phosphopeptide samples but also in phosphoproteomics SWATH-MS and DIA data from real biological samples, when compared to the other three library-free strategies. For phosphoproteomics diaPASEF data, Spectronaut's directDIA and the in silico-predicted library based on DIA-NN identify almost the same number of phosphopeptides as a project-specific DDA spectral library. However, only about 30% of the total phosphopeptides are commonly identified, suggesting that the library-free strategies for phospho-diaPASEF data need further improvement in terms of sensitivity. For ubiquitinomics diaPASEF data, the in silico-predicted library performs the best among the four workflows and detects ∼50% more K-GG peptides than a project-specific DDA spectral library. Our results demonstrate that Spectronaut's directDIA is suitable for the analysis of phosphoproteomics SWATH-MS and DIA MS data, while the in silico-predicted library based on DIA-NN shows substantial advantages for ubiquitinomics diaPASEF MS data.

Integrative genomic and transcriptomic analysis reveals genetic alterations associated with the early progression of follicular lymphoma.

Follicular lymphoma (FL), the most common indolent lymphoma, is a clinically and genetically heterogeneous disease. However, the prognostic value of driver gene mutations and copy number alterations has not been systematically assessed. Here, we analysed the clinical-biological features of 415 FL patients to identify variables associated with disease progression within 24 months of first-line therapy (POD24). Patients with B symptoms, elevated lactate dehydrogenase and ß2-microglobulin levels, unfavourable baseline haemoglobin levels, advanced stage, and high-risk FL International Prognostic Index (FLIPI) scores had an increased risk of POD24, with FLIPI being the most important factor in logistic regression. HIST1H1D, identified as a driver mutation, was correlated with POD24. Gains of 6p22.2 (HIST1H1D) and 18q21.33 (BCL2) and loss of 1p36.13 (NBPF1) predicted POD24 independent of FLIPI. Gene expression profiling of FL samples showed that the POD24 cohort was significantly enriched in the inflammatory response (mediated by interferon and tumour necrosis factor), cell cycle regulation (transcription, replication and proliferation) sets and PI3K-AKT-mTOR signalling. This result was further validated with transcriptome-wide information provided by RNA-seq at single-cell resolution. Our study, performed on a large cohort of FL patients, highlights the importance of distinctive genetic alterations and gene expression relevant to disease diagnosis and early progression.

Fast algorithm of single frequency-swept interferometry for the dynamic axial clearance measurement of high-speed rotating machinery.

Utilizing the periodicity of the rotating machinery, dynamic clearance measurement can be achieved with a single swept light source without any additional auxiliary devices, which has the advantages of simplicity and reliability. However, there is a shortcoming that previous algorithm is not fast enough to achieve real-time measurement when the machinery rotates at high speed. Aiming at this shortcoming, utilizing the correlation between mimic signal and measurement signal, combined with information for multiple periods, the fast algorithms and dynamic clearance corrected model were proposed. And the relationship between demodulation speed and cycle numbers was also discussed. Simulation was carried out to discuss the influence of different algorithm on the demodulation speed and accuracy. A test system was set up in the simulated environment for clearance measurement, and the results show that, the demodulation time of the proposed algorithm costs decreased dramatically, the speed has increased by about ten times, and the dynamic clearance measurement error is less than 2 µm.

Ultra-high sensitivity microwave-photonic sapphire fiber Fabry-Perot interferometry based on the Vernier effect.

The wavelength of microwave is longer than that of lightwave, causing the interferometric optical path difference induced by the measurand changes to be relatively smaller than that of microwave, which results in the lower sensitivity of sapphire fiber Fabry-Perot interferometer (FPI) sensor in microwave band. To improve sensitivity, a parallel FPI sensing system is constructed, in which a section of sapphire fiber connected to a single-mode transmission fiber is used as a sensing FPI, and a single-mode fiber (SMF) with a slightly different optical path from a sensing FPI is utilized as a reference FPI. By connecting two FPIs in parallel, Vernier effect is formed to improve sensitivity. The influence of relationship between the optical path difference of the reference FPI and the sensing FPI on the sensitivity amplification factor is analyzed based on the microwave interference spectrum of the parallel FPI. A section of sapphire fiber with the length of 8 cm is used as temperature sensor to construct high-temperature sensing system. The results demonstrate that the temperature sensitivity reaches about 2338.68kHz/°C, which is approximately 130 times higher than that of the sensing FPI alone. Furthermore, when the difference of optical path between the sensing FPI and the reference FPI is kept constant while the sensing FPI is unchanged, the amplification factor of the temperature sensitivity is approximately 2.64 times higher with longer length of the reference FPI compared to the situation with shorter length of the reference FPI.

Improved method to stratify lymphoma patients with risk of secondary central nervous system involvement: A multicenter retrospective analysis.

Secondary central nervous system (SCNS) involvement is an infrequent but universally fatal event in diffused large B-cell lymphoma. The occurrence rate of SCNS involvement is approximately 5% but comes with a poor prognosis ever after. However, existing risk models to predict the incidence and prognosis of these patients with SCNS involvement lack both efficiency and accuracy. Controversy has also been reported regarding which risk factor may best identify the population with a high CNS relapse rate. In this study, we retrospectively analyzed 831 patients with diffused large B-cell lymphoma, diagnosed between March 2008 and June 2018 in Tianjin Medical University Cancer Institute and Hospital, Beijing Cancer Hospital, and Cancer Hospital of The University of Chinese Academy of Science. Risk factors and nomogram were identified and established based on Fine and Gray's competing risk analysis. Among these patients, 55 (6.6%) of them eventually developed SCNS involvement. The 1- and 2-year incidence for SCNS involvement were 3.9% and 4.7%, respectively. The median time from de novo diagnosis to CNS relapse was 8 months, and the median overall survival of these patients was 28 months. Considering the competing mortality before SCNS involvement, Fine and Gray's competing risk model was performed to analyze the characteristics related to SCNS involvement, and identified risk factors as the multiple extranodal involvements, elevated LDH and AMC level, and the involvement of breast, adrenal gland/kidney, pulmonary and bone. Corresponding factors were integrated into the competing nomogram for SCNS involvement (c-index = 0.778). In conclusion, we present the first predictive nomogram to evaluate the risk to develop SCNS involvement in de novo DLBCL patients, which may help in both prognostic evaluation and clinical decision for this subgroup.

Realizing Ultrafast and Robust Sodium-Ion Storage of Iron Sulfide Enabled by Heteroatomic Doping and Regulable Interface Engineering.

Fe-based sulfides are a promising type of anode material for sodium-ion batteries (SIBs) due to their high theoretical capacities and affordability. However, these materials often suffer from issues such as capacity deterioration and poor conductivity during practical application. To address these challenges, an N-doped Fe7S8 anode with an N, S co-doped porous carbon framework (PPF-800) was synthesized using a template-assisted method. When serving as an anode for SIBs, it delivers a robust and ultrafast sodium storage performance, with a discharge capacity of 489 mAh g-1 after 500 cycles at 5 A g-1 and 371 mAh g-1 after 1000 cycles at 30 A g-1 in the ether-based electrolyte. This impressive performance is attributed to the combined influence of heteroatomic doping and adjustable interface engineering. The N, S co-doped carbon framework embedded with Fe7S8 nanoparticles effectively addresses the issues of volumetric expansion, reduces the impact of sodium polysulfides, improves intrinsic conductivity, and stimulates the dominant pseudocapacitive contribution (90.3% at 2 mV s-1). Moreover, the formation of a stable solid electrolyte interface (SEI) film by the effect of uniform pore structure in ether-based electrolyte produces a lower transfer resistance during the charge-discharge process, thereby boosting the rate performance of the electrode material. This work expands a facile strategy to optimize the electrochemical performance of other metal sulfides.

ACO:lossless quality score compression based on adaptive coding order.

BACKGROUND: With the rapid development of high-throughput sequencing technology, the cost of whole genome sequencing drops rapidly, which leads to an exponential growth of genome data. How to efficiently compress the DNA data generated by large-scale genome projects has become an important factor restricting the further development of the DNA sequencing industry. Although the compression of DNA bases has achieved significant improvement in recent years, the compression of quality score is still challenging. RESULTS: In this paper, by reinvestigating the inherent correlations between the quality score and the sequencing process, we propose a novel lossless quality score compressor based on adaptive coding order (ACO). The main objective of ACO is to traverse the quality score adaptively in the most correlative trajectory according to the sequencing process. By cooperating with the adaptive arithmetic coding and an improved in-context strategy, ACO achieves the state-of-the-art quality score compression performances with moderate complexity for the next-generation sequencing (NGS) data. CONCLUSIONS: The competence enables ACO to serve as a candidate tool for quality score compression, ACO has been employed by AVS(Audio Video coding Standard Workgroup of China) and is freely available at https://github.com/Yoniming/ACO.

A novel immune-related epigenetic signature based on the transcriptome for predicting the prognosis and therapeutic response of patients with diffuse large B-cell lymphoma.

Epigenetic modifications contribute to lymphomagenesis. Here, we performed an expression clustering analysis and identified two epigenetic-related clusters (EC1 and EC2). EC1 presented abundant TP53, MYD88, HIST1H1D, HIST1H1C, KMT2D and EZH2 mutations and an inferior prognosis. Pathways involved in the regulation of DNA methylation/demethylation, histone methyltransferase activity, and protein methyltransferase activity were significantly enriched in EC1. However, EC2 was frequently accompanied by B2M, CD70 and MEF2B mutations, which presented with enrichments in DNA damage repair, cytokine-mediated and B-cell activated immune signaling, increased levels of CD8+ T-, γδT- and T helper-cells, as well as immune scores and immunogenic cell death (ICD) modulators. According to the prediction, EC1 was more sensitive to vorinostat, serdemetan and navitoclax. However, ruxolitinib, cytarabine and CP466722 were more suitable treatments for EC2. The novel immune-related epigenetic signature exhibits promising clinical predictive value for diffuse large B-cell lymphoma (DLBCL), particularly for guiding epigenetic therapeutic regimens. R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone) based combination treatment regimens are suggested.

Tracking the evolution of untreated high-intermediate/high-risk diffuse large B-cell lymphoma by circulating tumour DNA.

Diffuse large B-cell lymphoma (DLBCL) is a highly heterogenous malignancy, early identification of patients for relapse remains challenging. The potential to non-invasively monitor tumour evolutionary dynamics of DLBCL needs to be further established. In the present study, 17 tumour biopsy and 38 plasma samples from 38 patients with high-intermediate/high-risk DLBCL were evaluated at baseline. Longitudinal blood samples were also collected during therapy. Circulating tumour DNA (ctDNA) was analysed using targeted sequencing based on a gene panel via a recently developed methodology, circulating single-molecule amplification and re-sequencing technology (cSMART). We found that the most frequently mutated genes were tumour protein p53 (TP53; 42·1%), histone-lysine N-methyltransferase 2D (KMT2D; 28·9%), caspase recruitment domain family member 11 (CARD11; 21·1%), cAMP response element-binding protein binding protein (CREBBP; 15·8%), ß2 -microglobulin (B2M; 15·8%), and tumour necrosis factor alpha-induced protein 3 (TNFAIP3; 15·8%). The mutation profiles between ctDNA and matched tumour tissue showed good concordance; however, more mutation sites were detected in ctDNA samples. Either TP53 or B2M mutations before treatment predicted poor prognosis. Analysis of dynamic blood samples confirmed the utility of ctDNA for the real-time assessment of treatment response and revealed that the increases in ctDNA levels and changes in KMT2D mutation status could be useful predictors of disease progression. Our present results suggest that ctDNA is a promising method for the detection of mutation spectrum and serves as a biomarker for disease monitoring and predicting clinical recurrence.

Learning to autofocus in whole slide imaging via physics-guided deep cascade networks.

Whole slide imaging (WSI), is an essential technology for digital pathology, the performance of which is primarily affected by the autofocusing process. Conventional autofocusing methods either are time-consuming or require additional hardware and thus are not compatible with the current WSI systems. In this paper, we propose an effective learning-based method for autofocusing in WSI, which can realize accurate autofocusing at high speed as well as without any optical hardware modifications. Our method is inspired by an observation that sample images captured by WSI have distinctive characteristics with respect to positive / negative defocus offsets, due to the asymmetry effect of optical aberrations. Based on this physical knowledge, we develop novel deep cascade networks to enhance autofocusing quality. Specifically, to handle the effect of optical aberrations, a binary classification network is tailored to distinguish sample images with positive / negative defocus. As such, samples within the same category share similar characteristics. It facilitates the followed refocusing network, which is designed to learn the mapping between the defocus image and defocus distance. Experimental results demonstrate that our method achieves superior autofocusing performance to other related methods.

Lensfree on-chip microscopy based on single-plane phase retrieval.

We propose a novel single-plane phase retrieval method to realize high-quality sample reconstruction for lensfree on-chip microscopy. In our method, complex wavefield reconstruction is modeled as a quadratic minimization problem, where total variation and joint denoising regularization are designed to keep a balance of artifact removal and resolution enhancement. In experiment, we built a 3D-printed field-portable platform to validate the imaging performance of our method, where resolution chart, dynamic target, transparent cell, polystyrene beads, and stained tissue sections are employed for the imaging test. Compared to state-of-the-art methods, our method eliminates image degradation and obtains a higher imaging resolution. Different from multi-wavelength or multi-height phase retrieval methods, our method only utilizes a single-frame intensity data record to accomplish high-fidelity reconstruction of different samples, which contributes a simple, robust, and data-efficient solution to design a resource-limited lensfree on-chip microscope. We believe that it will become a useful tool for telemedicine and point-of-care application.

A novel clinical immune-related prognostic model predicts the overall survival of mantle cell lymphoma.

The mantle cell lymphoma (MCL) International Prognostic Index (MIPI) and combined MIPI (MIPI-c) are commonly used for risk classification of MCL patients. However, these indexes lack immune-related parameters. The purpose of this study was to develop a novel prognostic model that integrated clinical and immune parameters. A total of 189 patients with newly diagnosed MCL from January 2010 to June 2020 were enrolled in our study. A nomogram and immune-related prognostic index (IRPI) were established to predict the overall survival (OS) of patients according to univariate and multivariate analyses. Discrimination and calibration were used to compare the prognostic performance of the IRPI, MIPI, and MIPI-c. External validation was performed based on validation dataset (n = 150) from two other centers. The results for the training dataset indicated that B symptoms, platelet count, B2M level, CD4+ T-cell count<26.7% and CD8+ T-cell count>44.2% were predictors for OS. All the prognostic factors were integrated into the nomogram. For the overlap of confidence intervals of each variable, we assigned one point for each factor. The IRPI categorized patients into three risk categories: a score of zero indicated low risk, a score of one or two indicated intermediate risk, and a score of ≥3 indicated high risk. The IRPI showed better discrimination and calibration power than the MIPI and MIPI-c in the training dataset and validation dataset. The novel IRPI is a refined risk stratification index and reflects the strong complementary prognostic effects between clinical and immune parameters in MCL.

Risk factors for POD24 in patients with previously untreated follicular lymphoma: a systematic review and meta-analysis.

Progression of disease within 24 months (POD24) is strongly associated with a poor outcome in patients with follicular lymphoma (FL). Our study aimed to identify the potential risk factors for POD24 in patients with FL. Medline, EMBASE and the Cochrane Library were systematically searched from the earliest record to September 2020. Studies investigating the prognostic factors for POD24 in patients with newly diagnosed grade 1-3a FL were included. Among 10,014 pieces of literature, a total of 90 studies investigating 82 risk factors were included for qualitative analysis. Meta-analyses were performed in 31 studies with 11 factors. Results showed that elevated sIL-2R, ß2m and LDH, total metabolic tumour volume > 510 cm3, vitamin D < 20 ng/mL, grade 3a and lymphoma-associated macrophages/high-power field ≥ 15 were significantly associated with an increased risk of POD24. No significant association was found between POD24 and the ALC/AMC ratio, sex, T effector signature or EZH2 genetic alteration. Additionally, minimal residual disease, Ki-67, PD-1 and TP53 were analysed narratively. Overall, this is the first study that comprehensively analysed the prognostic factors associated with POD24 in FL patients. We have confirmed the significance value of several common prognostic factors as well as others not commonly included in clinical study, helping to construct an integrated and more efficient model.