Fluorescence biosensing is a widely employed technique in the field of life sciences due to its ability to visualize multiplexed physiological and pathological information with rapid feedback, high sensitivity, and high spatiotemporal resolution ^[1][2][3][1,2,3]. Multichannel biosensing is particularly significant for the development of biomedicine in both fundamental research and clinical practice ^[4][5][4,5]. It is capable of monitoring multiple processes or quantitatively/qualitatively detecting biomarkers, providing accurate and valuable diagnostic information on tissue, cellular, and molecular features ^[6][7][6,7]. Nevertheless, multichannel fluorescence biosensing in the visible (400–700 nm) and first near-infrared (NIR-I, 700–950 nm) region encounters poor photon penetration depth in mammalian tissues due to the substantial photon scattering/absorption and the interferential autofluorescence ^[8][9][10][8,9,10]. In contrast, fluorescence imaging in the second near-infrared window (NIR-II, 950–1700 nm) can offer deeper penetration depth and better clarity owing to the suppressed photon scattering (us’∝λ-α; u: scattering coefficient; λ: wavelength; and α: constant) and reduced tissue autofluorescence (inversely proportional to wavelength) compared with the visible and NIR-I regions ^{[11][12][13][14]}[11,12,13,14]. Thus, the advancement of new NIR-II fluorescent probes is necessary to support the development of deep-tissue and multichannel fluorescence biosensing.

NIR-II fluorophores, mainly including organic dyes ^[15][16][15,16], quantum dots (QDs) ^[17][18][17,18], and rare-earth-based nanoparticles (RENPs) ^[19][20][19,20], have been developed for multichannel fluorescence imaging using spectra-domain, lifetime-domain, and fluorescence-lifetime modes. The fabrication of these fluorophores, which exhibit nonoverlap emissions–excitations or distinguishable fluorescence lifetimes, is crucial for enabling multichannel biosensing applications ^[21][22][21,22]. To facilitate the qualitative and quantitative analysis of niduses, organic dyes may be designed to control their chemical groups and conjugation units, thereby regulating their fluorescence emissions–excitations and responsiveness to pathological environments ^[23][24][23,24]. QDs may be tuned via the regulation of their size and chemical structure (e.g., core–shell structure and heteroatom doping) to alter their fluorescence emissions ^[17][25][17,25]. RENPs can be modified by doping with various sensitizers/activators and by altering their core–shell structures (e.g., interlayer and shell thickness), resulting in changes in their fluorescence emissions–excitations and lifetimes ^[19][26][19,26]. These fluorescent probes, possessing unique fluorescence and lifetime properties, can be utilized as a versatile toolbox to fulfill the demands of multiscale and multichannel biosensing in a variety of scenarios.

The implementation of in vivo multichannel fluorescence biosensing requires a meticulous selection of nonoverlap excitation–emission wavelengths for the use of NIR-II fluorophores. To achieve simultaneous visualization/tracking of multiple analytes at the tissue or cellular level, fluorophores with varying excitation–emission properties are necessary to bind specific targets. On the other hand, in order to quantitatively measure analytes at the molecular level, ratiometric fluorophores with multiple excitation–emission wavelengths are typically required to enable selective responsiveness. The rational designs and bioapplications of these NIR-II fluorophores will be elucidated in-depth in subsequent sections.

In order to facilitate the visualization of multiple pathological or physiological parameters at the tissue or cellular level, it is essential to use fluorophores with nonoverlap excitation–emission spectra that enable to simultaneously tag and visualize them in the NIR-II imaging window. Recently, Cosco et al. tuned the absorption properties of flavylium polymethine dyes utilizing flavylium heterocycles, which allowed for the real-time excitation multiplexing of living mice at the tissue level using NIR-II emitters (MeOFlav7: 980 nm excitation; JuloFlav7: 1064 nm excitation) ^[27][36]. However, the long-term retention of these molecules has the potential to induce toxicity in major organs. To overcome this limitation, Yao et al. designed a series of highly efficient renal clearance and long blood circulation aza-boron-dipyrromethene (aza-BODIPY) NIR-II macromolecular fluorophores (FBP790: 730 nm excitation, 950 nm long-pass optical filter; FBP1025: 980 nm excitation, 1200 nm long-pass optical filter) and applied them for excitation multiplexed imaging of tumors and surrounding vessels ^[28][47]. Nevertheless, these NIR-II organic fluorophores are restricted to the emission wavelength range of 1000–1500 nm, thereby limiting their utility in achieving high-resolution NIR-IIb fluorescence imaging (1500–1700 nm) with millimeter-scale penetration depth and micron-level resolution ^[29][30][31][48,49,50]. Although AIEgens (DCBT dots) with multiphoton absorption could obtain fine structures of brain vasculatures of mice and macaque by using three-photon fluorescence bioimaging with NIR-IIb excitations (1550 nm), their progress in the field of excitation–emission multiplexed biosensing remains sluggish ^[32][33][34][51,52,53].

To date, only a handful of NIR-IIb fluorophores have been developed for multichannel fluorescence biosensing, including rare-earth-based nanoprobes and colloidal QDs ^[35][36][54,55]. Given the intrinsic location of 4f energy levels of lanthanides, Er(III) is the mainstream activator generating NIR-IIb emission (around 1530 nm) under varied excitation sources (650 nm, 808 nm, and 980 nm) ^{[37][38][39][40]}[32,56,57,58]. To expand the range of small-molecule lanthanide complexes for biosensing, Wang et al. developed a hybrid Er(III)-bacteriochlorin complex (EB766) with 760 nm excitation and bright luminescence at 1530 nm ^[41][59]. Compared with typical co-doped lanthanide nanoparticles (ErNPs: NaYF₄:20%Yb,2%Er@NaYF₄), the nonoverlap excitations of EB766 (760 nm) and ErNPs (980 nm) facilitate robust NIR-IIb multiplexed biosensing for deciphering multiple tissue structures (such as lymph nodes and lymph vessels in jaw and footpads) with superior signal-to-background ratios (SNRs: 3.92–22.79). Alongside lanthanides, colloidal QDs, such as PbS QDs with a narrow bandgap and large Bohr radius, offer tunable emission coverage of the NIR-II region via size control ^[42][43][60,61]. By synthesizing NIR-IIa and NIR-IIb fluorescent PbS/CdS core–shell QDs functionalized with Gr-1 and CD-1b antibodies, respectively, Yu et al. exploited this emission multiplexing strategy for in vivo colocalization of myeloid-derived suppressor cells (MDSCs). By connecting the unique emission properties of these QDs to the targeted MDSCs, this approach holds significant potential for enhancing immunotherapy [21]. Overall, these results demonstrate that the excitation–emission multiplexing approach is highly suitable for a real-time diagnosis of tissue lesions and therapeutic feedback at the cellular level.

In addition to multichannel biosensing utilizing multiple fluorophores, fluorophores with the capacity to emit or excite at different wavelengths have the potential to enable the quantitative measurement of pathological parameters at the molecular level. For this purpose, the fluorophores must possess an emission–excitation pattern that responds solely to the analyte (such as reactive oxygen species, enzymes, and pH), while the other remains inert and acts as a reference signal. To achieve this, Lan et al. designed a series of responsive NIR-II probes (enzyme-responsive Rap-N, ROS-responsive Rap-R, and pH-responsive Py-H; ratiometric emission: 945 nm/1010 nm) based on the Py-2 molecular platform for ratiometric fluorescence biosensing [24]. These single-component NIR-II dyes have the potential to selectively visualize and quantitatively measure enzymes and small molecules in living mice with significant ratiometric fluorescence signals (Fl900LP/Fl1000LP). Additionally, NIR-II organic nanosensors for ratiometric fluorescence imaging can be designed based on intramolecular Förster resonance energy transfer (FRET). To create the NIR-II ratiometric fluorescent dyes, Yu et al. covalently linked an asymmetric aza-BODIPY with an ONOO⁻-responsive meso-thiocyanine, known as the aBOP-IR1110 (ratiometric emission: 950 nm/1110 nm). This process results in ONOO⁻-altered intramolecular FRET, which generates a linear ratiometric response (Fl950LP/Fl1100LP) ^[44][62]. The aBOP-IR1110 can withstand biological media, thereby preventing spectral shift and distortion and facilitating the dynamic monitoring of oxidative stress in traumatic brain injury and evaluating therapeutic efficiency with high in vivo sensing accuracy. However, the currently available organic ratiometric fluorophores have a limited emission wavelength that falls short of 1500 nm, along with a small Stokes shift, thereby limiting their usability in deep-tissue biosensing.

Hybrid ratiometric nanosensors offer a variety of options for emission wavelength and responsive substances, thus enabling quantitative in vivo biosensing. Sun et al. developed a NIR-II ratiometric nanocomposite by coupling the dual-emission (1060 nm and 1525 nm) RENPs (NaErF₄@NaYF₄@NaYF₄:10%Nd@NaYF₄) with ONOO⁻-responsive A1094 organic dyes (absorbance:1094 nm). The FRET between them allowed for the rapid and sensitive in vivo detection of ONOO⁻ levels ^[45][63]. Since ROS/RNS levels are highly correlated with the activation and viability of immune cells, in vivo molecular analysis can provide valuable feedback during cancer immunotherapy ^[46][47][65,66]. Liao et al. developed a NIR-II ratiometric nanocomposite by coating dual-excitation (ex: 808 nm and 980 nm; em: 1525 nm) RENPs with IR786 dyes. The degradation of ROS-responsive dyes solely activates the 808 nm excitation channel through the absorption competition-induced emission effect (ACIE) ^[48][64]. Therefore, the hybrid nanosensor can evaluate the cell viability of natural killer (NK) cells by measuring the excess generation of ROS, while simultaneously tracking the NK cells via the stable signal excited at 980 nm. Such intracellular ROS-induced ratiometric NIR-II fluorescence biosensing has the potential to pave the way for in vivo companion diagnostics in cancer immunotherapy.

In the field of oncology, mapping the heterogeneity of primary and metastatic tumors is vital for precision medicine [4]. Detecting the metastasis of cancer cells and the migration of immune cells is essential from a fundamental research perspective, as it can help identify the most effective therapies ^[49][50][67,68]. To study cellular behavior in live animals, intravital microscopic multiplexing is essential ^[51][69]. Wang et al. developed a cell tracking probe (CT1530) for intracellular delivery, comprising the cell-penetrating peptide HIV-TAT-conjugated EB766–bovine serum albumin (BSA) complex ^[41][59]. Using fluorescence signals (green, 1100–1300 nm) from Cy7.5 phospholipid micelles to highlight the cerebrovascular system, the researchers found that cancer cells were arrested in vessel bifurcations through physical occlusion via the nonoverlap fluorescence (red, 1400–1600 nm) from CT1530-labeled 143B cells. On the other hand, evaluating the infiltration of immune cells into tumors is crucial for determining therapeutic efficacy. Hao et al. developed a NIR-II emission multiplexing strategy to visualize the recruitment of NK cells into the tumor via the programmable administration of Ag₂Se-QD-based nanodrug (em: 1350 nm) containing SDF-1α (stromal-cell-derived factor-1α, the chemokine of NK-92 cells) and Tat-Ag₂S QDs-labeled NK-92 cells (em: 1050 nm) ^[52][70]. This multichannel biosensing strategy has significant implications not only for precision medicine but also for individual therapeutic schedules and companion diagnostics ^[4][53][4,71]

In clinical practice, a first-in-human study has demonstrated the potential of intraoperative NIR-II fluorescence imaging in guiding liver tumor surgery [5]. In addition to primary tumors theranostics, detecting and monitoring cancer metastasis is critical for tumor staging, therapeutic decision making, and prognosis. For instance, in the case of breast cancer, multiorgan metastasis often leads to a median survival rate of fewer than two years for patients. Lymph nodes (LNs) are frequently the primary site of tumor cell dissemination, which can disrupt the immune microenvironment ^[54][74]. To enable precise surgical resection, Tian et al. developed an emission multiplexing approach that simultaneously labels the metastatic tumor (IR-FD dye) and tumor metastatic proximal LNs (PbS/CdS QDs) ^[55][72]. Meanwhile, diagnosing allograft rejection to improve the immune management of transplant recipients in the early stages is also essential, as transplantation can cause severe postoperative complications ^[56][57][75,76]. Chen et al. developed a responsive NIR-II fluorescent nanosensor by linking ErNPs (NaErF₄@NaYF₄: 980/808 nm excitation; 1550 nm emission) with ZW800 dye, allowing for the ratiometric biosensing of granzyme B, which is overexpressed in recipients’ T cells during the onset of allograft rejection, in contrast to the gold-standard biopsy ^[58][73]. This strategy could also be applied to in situ monitoring of tissue regeneration. Pei et al. integrated 3D-printed bioactive glass scaffolds with a responsive NIR-II fluorescent nanosensor, enabling in situ monitoring of early inflammation, angiogenesis, and implant degradation during mouse skull repair ^[59][77]. These results illustrate the potential of spectral-domain fluorescence multiplexing in clinical translation.