用于指导基于自然杀伤细胞疗法的实体瘤适应症的多模态框架。
全文（开放获取） · 中文译文
摘要 基于T细胞的免疫疗法在血液系统恶性肿瘤中取得了显著成功，但在实体瘤中仍然有限。自然杀伤（NK）细胞提供了一种有吸引力的互补方法，因为它们能够独立于新抗原呈递和MHC I类分子表达来识别和杀伤肿瘤细胞。本研究整合了多组学分析、临床试验数据和转录组分析，以系统地识别最有可能从基于NK细胞的疗法中获益的实体瘤适应证。免疫组织化学分析显示，NK细胞浸润多种肿瘤类型，其中在肾脏肿瘤中丰度最高。对177项临床试验的系统评价发现，不同实体瘤适应证之间的疗效存在差异，没有明显更优的肿瘤类型或干预措施。对超过10,000份TCGA样本的转录组分析评估了NK细胞浸润、细胞毒性适应性和肿瘤微环境组成，指出肾细胞癌、肺腺癌和肺鳞状细胞癌、宫颈癌和宫颈内膜癌以及间皮瘤是未来临床评估的有希望的候选类型。总之，这些发现提供了一个多模式框架，用于对肿瘤适应证进行优先排序，并改进未来NK细胞临床试验的合理设计。 基于T细胞的免疫疗法在治疗血液系统恶性肿瘤方面已取得了不可否认的临床成功 [, , , ]，而在实体瘤中观察到的结果则逊色得多，最近仅有少数基于T细胞的产物获得FDA批准：lifileucel（自体肿瘤浸润淋巴细胞）和afami-cel（TCR工程自体T细胞）分别用于晚期黑色素瘤和滑膜肉瘤 [, ]，以及T细胞衔接双特异性抗体tebentafusp（抗gp100/HLA-A*02:01 × CD3）用于葡萄膜黑色素瘤 [] 和tarlatamab（抗DLL3 × CD3）用于广泛期小细胞肺癌 []。此外，获得持久缓解的患者数量仍然有限，促使需要补充治疗方法。自然杀伤（NK）细胞在肿瘤控制中发挥多方面的作用，即直接细胞毒性和促炎细胞因子的分泌。此外，它们独立于新抗原呈递识别应激细胞，并对通过丢失MHC I类分子表达而获得对T细胞免疫耐药性的肿瘤表现出增强的细胞毒性 [, , ]。 在发现NK细胞同种异体反应性可预测MHC半相合造血细胞移植（HCT）背景下急性髓系白血病（AML）患者的无进展生存期后，NK细胞被评估为血液癌症的细胞疗法[]。在第一项同种异体NK细胞的1期临床试验中，Miller及其同事在淋巴细胞清除化疗后使用来自MHC半相合供体的富集常规NK（cNK）细胞产品，并辅以rhIL-2支持[]。他们表明，约25%的复发/难治性AML患者安全地达到了复合完全缓解（CR），无需进行HCT，尽管缓解仅持续数月。该方法的扩展包括去除Treg；改变IL-2给药方案也安全地诱导了缓解，其缓解率高于未去除Treg的方案（53%对21%），33%的患者在6个月时仍处于持续缓解状态[]。值得注意的是，血液中的NK细胞扩增与临床缓解显著相关。在一种补充方法中，一项1期临床试验在过继转移纯化供体NK细胞之前使用IL-12、IL-15和IL-18预激活以诱导记忆样程序[]。这使复发/难治性AML患者的复合CR率达到47%，缓解持续时间为6–9个月。许多同种异体非工程化血液NK细胞的试验显示，与历史对照相比，早期临床试验的CR率和/或无进展生存期均有改善，凸显了NK细胞疗法的前景[, ]，该领域的共识是非工程化NK细胞可在临床中产生…… 对血癌患者，尤其是急性髓系白血病（AML）患者有益。然而，在AML中尚未开展支持同种异体NK细胞疗法广泛应用所需的大型前瞻性随机试验，以支持其在AML及相关疾病中的广泛使用。这很可能是因为商业化挑战以及NK细胞工程策略的出现，这些策略为增强NK细胞对癌症的反应提供了新方法。 随着通过细胞工程将CAR导入NK细胞取得成功，大多数利用NK细胞的临床试验正在改造NK细胞，使其表达合成激活受体，从而提供额外的激活信号，并在某些情况下提供细胞因子支持。临床上最早测试的CD19‐CAR工程化脐带血来源NK细胞显示出令人鼓舞的早期CR率，并证明异体CD19‐CAR NK细胞可在淋巴瘤患者中产生持久CR，12个月PFS为32%。同样，由iPSC分化而来的CD19‐CAR NK细胞在一组异质性B细胞淋巴瘤患者中显示出临床活性，包括那些在接受自体CAR T细胞治疗后出现疾病进展的患者。基于这些早期成功，大量正在进行的临床试验正在探索多种CAR靶点和策略，以增强工程化NK细胞的活性，包括靶向实体瘤。然而，基于NK细胞的疗法在实体瘤中的有效性仍不充分，主要原因是目前对肿瘤浸润NK细胞了解不完整，尤其是对其浸润或排除的调控机制、其表型和功能异质性，以及其在肿瘤微环境中的功能失调机制的认识不足。 鉴于实体瘤的显著异质性，既体现在免疫细胞浸润方面，也体现在肿瘤细胞内在特征方面，例如激活性或抑制性NK细胞配体的表达以及调节NK细胞募集的趋化因子分泌 [, ]，我们旨在识别最有可能从增强NK细胞活性的疗法中获益的实体瘤类型。聚焦这些实体瘤适应症将提高基于NK细胞策略的转化相关性，并增强实现有意义临床获益的概率。我们的策略整合了大规模基因组数据集分析与来自免疫组织化学和评估NK细胞输注的临床试验的见解，从而能够全面理解影响治疗疗效的因素。 NK细胞浸润多种实体瘤类型，并在肾肿瘤中富集 NKp46是组织中使用的经典NK细胞标志物 []。我们分析了Gautier等人 [] 生成的免疫组织化学数据，该数据报告了使用抗NKp46染色检测人类实体瘤（肺、胰腺、胃、肾、结肠、头颈和肝）样本中NK细胞的存在情况，以确定哪些适应症以NK细胞丰度更高为特征（图 ）。在所有分析的肿瘤适应症中均可检测到NK细胞，其中肾肿瘤的检出率最高。 NK细胞在实体瘤适应证中的分布。商品化的福尔马林固定石蜡包埋（FFPE）肿瘤组织芯片（来自US Biomax的肺（#LC10012a）、胰腺（#PA1001b）、胃（#ST1003a）、肾（#HKid‐CRCC150CS‐02）、结肠（#BC051110b）、头颈部（#HN802b）和肝（#BC03119a））使用抗NKp46抗体（克隆8E5B，Innate Pharma）进行染色，并使用抗小鼠/兔IgG检测系统（Ultraview universal DAB检测试剂盒，Ventana）进行一抗检测和扩增[28]。黑色柱状图代表检测到至少一个NKp46+细胞的肿瘤样本阳性百分比；数字代表每种肿瘤适应证检测的样本数量。 NK细胞疗法在实体瘤类型中的疗效 为了分析NK细胞输注在哪些实体瘤中显示出最显著的临床结果，我们在ClinicalTrials.gov数据库中进行检索，查找自2000年1月1日至2025年6月30日启动的所有涉及NK细胞输注治疗实体瘤的临床研究。检索策略采用的术语为：condition‐tumor、intervention‐NK cells、status‐all studies。共识别出807项试验，其中经人工筛选与实体瘤相关且将NK细胞作为干预措施后保留177项。其中，83项临床试验探索NK细胞作为单药使用，而在另外94项研究中，NK细胞与化疗、促炎细胞因子（主要为IL‐2或IL‐15）或抗体（靶向肿瘤抗原或免疫检查点抑制剂）联合使用。在近三分之二的研究中（177项中的114项），NK细胞未经基因修饰（自体或异体），而177项试验中的53项使用CAR‐NK细胞，仅10项使用NK‐92细胞系。49项研究的疗效结果可在Beacon数据库（data.beacon‐intelligence.com）中找到，并以总缓解率（ORR）表示于图。表总结了每项临床试验的干预措施。报告疗效结果的临床试验数量对于每种肿瘤适应证仍然有限。此外，各研究之间观察到的临床缓解存在显著差异，可能反映了NK细胞输注前接受过的既往治疗线数的差异。因此，没有任何单一的肿瘤适应证或特定干预措施能够 so f。 应被明确确定为展示出更优的临床获益。 实体瘤患者输注NK细胞的临床疗效。临床活性通过总缓解率（ORR）评估，总缓解率定义为所有可评估患者中达到完全缓解或部分缓解的患者比例。每个气泡代表一项临床试验，气泡大小代表可评估患者的数量。 基于转录组数据的肿瘤适应症聚类，以识别功能性NK细胞富集的肿瘤。 为了进一步探索最有可能对NK细胞疗法产生应答的实体瘤情况，我们分析了TCGA数据库中31种不同类型癌症、超过10,000个肿瘤样本的可用RNA-seq数据集，以评估NK细胞浸润、NK细胞适应度及微环境因素（表 ）。选择由NK细胞显著过表达的信息性基因（CHST12、KLRF1、GNLY、KLRC1、IL18RAP、KLRD1、IL2RB、CST7、XCL2、CD244、PRF1、NCR1、KIR2DL1、KIR2DL3、CD160和KLRC3）[]，用以推断所分析肿瘤样本中这些细胞类型的丰度。通过热图分析，出现了一组以NK细胞高浸润为特征的肿瘤（图 ）。随后，采用活化性NK细胞受体NKp30、CD16、KLRF1、CD160、DNAM1、KLRK1、CD244以及细胞毒性分子穿孔素和颗粒酶B的表达，来识别含有具有最高细胞毒性潜能的NK细胞的肿瘤[]（图 ）。在肿瘤微环境中，活化性和抑制性NK细胞受体的配体平衡，连同免疫抑制性可溶性因子（如TGF-β、PGE2和IL-10），塑造了NK细胞功能[, ]。对这些分子表达的分析（图 ）划分出一个具有有利于NK细胞介导裂解的微环境特征的肿瘤亚群，包括肾透明细胞癌、肺腺癌和肺鳞状细胞癌、间皮瘤、胰腺腺癌、宫颈癌和宫颈内膜癌、头颈部鳞状细胞癌、膀胱尿路上皮癌和肉瘤。Notab ly，肺、肾、宫颈和宫颈内膜癌以及间皮瘤也显示出功能性 NK 细胞的大量浸润（图 ）。总的来说，这些发现表明肾细胞癌、肺腺癌和鳞状细胞癌、宫颈癌和宫颈内膜癌以及间皮瘤可能是基于 NK 细胞的疗法的有前途的候选类型（图 ）。然而，应该指出，这些结论基于批量转录组分析，这些分析本质上受到肿瘤样本异质性的限制，包括免疫细胞浸润、肿瘤分期和既往治疗等方面。 通过转录组数据集分析进行聚类以选择实体瘤适应症。通过对TCGA数据库中31种实体瘤类型的RNA‐seq数据集进行热图分析，评估指示NK细胞存在和适应性、激活性NK细胞受体配体以及NK细胞抑制性分子的基因表达，提出了最有可能对基于NK细胞的疗法产生反应的肿瘤适应症。首先，分析了由NK细胞显著过表达的基因（CHST12、KLRF1、GNLY、KLRC1、IL18RAP、KLRD1、IL2RB、CST7、XCL2、CD244、PRF1、NCR1、KIR2DL1、KIR2DL3、CD160、KLRC3）[]以及激活性NK细胞受体NKp30、CD16、KLRF1、CD160、DNAM1、KLRK1、CD244和细胞毒性分子穿孔素和颗粒酶B的表达，以鉴定细胞毒性NK细胞的存在（左）。然后，评估肿瘤微环境中NK细胞调节因子的阳性或阴性存在（激活性与抑制性NK细胞受体的配体以及可溶性因子，如TGF‐β、PGE2和IL‐10），以检测表现出有利于NK细胞介导裂解的微环境特征的肿瘤（右）。通过这两种分析均鉴定出的肿瘤类型以粗体突出显示。a，腺癌；cc，细胞癌；c，癌。 以瘤内 NK 细胞高浸润和适应性为特征的适应症可能特别适合利用体内内源性 NK 细胞的治疗策略（例如抗体或免疫细胞因子）。相比之下，在显示活化配体有利表达模式且抑制信号有限的适应症中，无论是否经过基因工程改造，过继性 NK 细胞输注方法都可能具有前景。对介导 NK 细胞在肿瘤中募集的趋化因子，即 CX3CL1、RARRES2（Chemerin）、CCL5、CXCL9、CXCL10 和 CXCL14 [, , , , ] 的表达也进行了分析，但在不同肿瘤适应症之间未观察到重大差异。绝大多数实体瘤似乎表达吸引 NK 细胞的趋化因子；然而，NK 细胞仅占肿瘤浸润淋巴细胞的一小部分。潜在的解释包括未达到所需阈值、限制 NK 细胞浸润的物理屏障，或缺乏维持 NK 细胞活力所需的生长因子 [, , ]。 结论 总的来说，这些数据为基于自然杀伤（NK）细胞的实体肿瘤适应症的合理优先级排序提供了一个整合性的多组学框架。通过系统整合NK细胞浸润的免疫组织化学证据、临床试验疗效数据、转录组谱分析和肿瘤微环境特征分析，我们确定肾透明细胞癌、肺腺癌和鳞状细胞癌、宫颈癌和宫颈内膜癌以及间皮瘤是呈现最有利于NK细胞介导抗肿瘤活性的生物学特征的肿瘤类型。这些适应症的特征是具有有意义的NK细胞浸润，以及肿瘤微环境中激活配体与有限的免疫抑制信号之间具有有利的平衡。 临床试验格局分析突显了当前NK细胞治疗领域在实体肿瘤中的异质性和碎片化状态，目前尚无单一干预措施或适应症显示出明确无疑的优越性。这些发现可能反映了患者既往治疗史的异质性，以及各项研究中可评估患者数量通常有限的情况。这强调了以生物标志物驱动的患者选择和标准化终点指标的迫切需求，以便进行有意义的跨研究比较。我们的转录组分析开始通过提出客观的、数据驱动的、超越经验观察的适应症选择标准来弥补这一空白。 综上所述，这些发现具有直接的转化意义。它们为将NK细胞临床开发聚焦于一组明确的实体瘤适应症提供了基于生物学依据的理由，从而提高了临床成功的概率，并加速了获得监管批准的路径。随着NK细胞工程不断取得进展——包括日益复杂的CAR构建体、细胞因子武装，以及克服微环境免疫抑制的策略——本文提出的框架为将这些创新与最可能发挥效力的实体瘤背景相结合提供了路线图。在专门的富集生物标志物的临床试验中对这些发现进行前瞻性验证，对于将这些多模态证据转化为改善的患者结局至关重要。 作者贡献 L.C.、V.P.、T.A.F.和E.V.构思了所呈现的数据。L.C.进行了免疫组织化学数据分析并审阅了临床试验结果。L.C.和Q.B.分析并解读了RNA-seq数据。L.C.、V.P.、Q.B.、T.A.F.和E.V.讨论了结果并审阅了手稿。 利益冲突 作者声明无利益冲突。 支持信息 支持文件1：eji70283‐sup‐0001‐SuppMat.pdf。 支持文件2：eji70283‐sup‐0002‐Table S1.xlsx。 支持文件3：eji70283‐sup‐0003‐Table S2.xlsx。
全文（开放获取） · 英文原文
【Abstract】 T cell-based immunotherapies have achieved remarkable success in hematological malignancies but remain limited in solid tumors. Natural killer (NK) cells offer a compelling complementary approach, given their ability to recognize and kill tumor cells independently of neoantigen presentation and MHC class I expression. This study integrates multiomic analyses, clinical trial data, and transcriptomic analyses to systematically identify solid tumor indications most likely to benefit from NK cell-based therapies. Immunohistochemical analysis revealed that NK cells infiltrate multiple tumor types, with the highest abundance in kidney tumors. Systematic review of 177 clinical trials identified variable efficacy across solid tumor indications, without a clearly superior tumor type or intervention. Transcriptomic profiling of over 10,000 TCGA samples assessed NK cell infiltration, cytotoxic fitness, and tumor microenvironment composition, highlighting kidney renal cell carcinoma, lung adenocarcinoma and squamous cell carcinoma, cervical and endocervical cancers, and mesothelioma as promising candidates for future clinical evaluation. Together, these findings provide a multimodal framework to prioritize tumor indications and improve the rational design of future NK cell clinical trials. 【Introduction】 T cell‐based immunotherapies have yielded undeniable clinical success in the treatment of hematological malignancies [, , , ] while much more modest results were observed in solid tumors, with only a handful of T cell‐based products recently approved by the FDA: lifileucel (autologous tumor‐infiltrating lymphocytes) and afami‐cel (TCR‐engineered autologous T cells) in advanced melanoma and synovial sarcoma [, ], respectively, as well as the T cell‐engaging bispecifics tebentafusp (anti‐gp100/HLA‐A*02:01 × CD3) in uveal melanoma [] and tarlatamab (anti‐DLL3 × CD3) in extensive‐stage small cell lung cancer []. Moreover, the number of patients who experience durable responses is still limited, prompting the need for complementary treatment approaches. Natural killer (NK) cells play multifaceted roles in tumor control, namely direct cytotoxicity and the secretion of proinflammatory cytokines. Moreover, they recognize stressed cells independently of neoantigen presentation and exhibit increased cytotoxicity toward tumors that acquire resistance to T cell immunity via loss of MHC class I expression [, , ]. NK cells were evaluated as a cellular therapy for blood cancers after the discovery that NK cell alloreactivity predicted progression‐free survival for acute myeloid leukemia (AML) patients in the setting of MHC‐haploidentical hematopoietic cell transplant (HCT) []. In the first phase 1 clinical trial of allogeneic NK cells, Miller and colleagues utilized enriched conventional NK (cNK) cell products from MHC‐haploidentical donors following lymphodepleting chemotherapy and support with rhIL‐2 []. They showed that ∼25% of rel/ref AML patients safely achieved a composite complete remission (CR) without the need for an HCT, albeit durable for only a few months. Expansions of this approach incorporated Treg depletion; altered IL‐2 schedules also safely induced remissions at a higher rate than without Treg depletion (53% vs. 21%), and 33% of patients had ongoing remissions at 6 months []. Notably, NK cell expansion in the blood was significantly associated with clinical response. In a complementary approach, a phase 1 clinical trial utilized IL‐12, IL‐15, and IL‐18 pre‐activation to induce memory‐like programs prior to adoptive transfer of purified donor NK cells []. This resulted in a composite CR rate of 47% in rel/ref AML patients, with durations of remission lasting 6–9 months. Numerous trials of allogeneic nonengineered blood NK cells have shown improvements in CR rate and/or PFS compared with historical controls in early‐phase clinical trials, highlighting the promise of NK cell therapies [, ], with a consensus in the field that nonengineered NK cells can result in clinical benefits to patients with blood cancers, especially AML. However, large prospective randomized trials required for broad adoption of allogeneic NK cell therapy have not been performed in AML to support their broad use in AML and related diseases. This is most likely due to the challenges of commercialization and the advent of NK cell engineering strategies that provide new approaches to enhance NK cell responses to cancer. With the success of cellular engineering to introduce CAR into NK cells [], most clinical trials utilizing NK cells are modifying NK cells to express synthetic activating receptors, thereby providing additional activating signals and, in some cases, cytokine support []. The first CD19‐CAR‐engineered cord blood‐derived NK cells tested in the clinic had a promising early CR rate [], and demonstrated that allogeneic CD19‐CAR NK cells could produce durable CRs in lymphoma patients with a 12‐month PFS of 32% []. Similarly, CD19‐CAR NK cells differentiated from iPSCs demonstrated clinical activity in a heterogeneous set of B cell lymphoma patients, including those that had progressed after autologous CAR T cells []. Based on these early successes, numerous ongoing clinical trials are exploring a variety of CAR targets and strategies to enhance the engineered NK cell activity, including targeting solid tumors [, ]. However, the effectiveness of NK cell‐based therapies in solid tumors remains inadequate, largely due to an incomplete understanding of tumor‐infiltrating NK cells, especially the mechanisms governing their infiltration or exclusion, their phenotypic and functional heterogeneity, and their functional dysregulation within the tumor microenvironment. Given the pronounced heterogeneity of solid tumors, both in immune cell infiltration and in tumor cell intrinsic features such as the expression of activating or inhibitory NK cell ligands and the secretion of chemokines that regulate NK cell recruitment [, ], we aimed to identify solid tumor types most likely to benefit from therapies that potentiate NK cell activity. Focusing on these solid tumor indications will refine the translational relevance of NK cell‐based strategies and enhance the probability of achieving meaningful clinical benefit. Our strategy integrates large‐scale genomic dataset analyses with insights derived from immunohistochemistry and clinical trials evaluating NK cell infusions, enabling a comprehensive understanding of factors influencing therapeutic efficacy. 【NK Cells Infiltrate Multiple Solid Tumor Types and Are Enriched in Renal Tumors】 NKp46 is the canonical NK cell marker used in tissue sections []. We analyzed immunohistochemistry data generated by Gautier et al. [] reporting the presence of NK cells in samples from human solid tumors (lung, pancreas, gastric, kidney, colon, head and neck, and liver) using anti‐NKp46 staining to identify which indications are characterized by a greater abundance of NK cells (Figure ). NK cells were detectable across all tumor indications analyzed, with the highest detection rate observed in kidney tumors. NK cell distribution across solid tumor indications. Commercially available formalin‐fixed paraffin‐embedded (FFPE) tumor tissue microarrays (lung (#LC10012a), pancreas (#PA1001b), gastric (#ST1003a), kidney (#HKid‐CRCC150CS‐02), colon (#BC051110b), head and neck (#HN802b) and liver (#BC03119a) from US Biomax) were stained with an anti‐NKp46 Ab (clone 8E5B, Innate Pharma) and anti‐mouse/rabbit. IgG detection system (Ultraview universal DAB detection kit, Ventana) was used for primary Ab detection and amplification [28]. Black bars represent the percentage of tumor samples positive where at least one NKp46+ cell was detected; numbers represent the number of samples tested for each tumor indication. 【Efficacy of NK Cell Therapy across Solid Tumor Types】 To analyze in which solid tumors NK cell infusions showed the most significant clinical results, we conducted a search on the ClinicalTrials.gov database looking for all clinical studies involving NK cell infusions in solid tumors started from January first, 2000, to June 30, 2025. The search strategy applied the terms: condition‐tumor, intervention‐NK cells, status‐all studies. 807 trials were identified, of which 177 were retained after manual screening for relevance to solid tumors and the use of NK cells as an intervention. Among them, 83 clinical trials explored the use of NK cells as a single agent, while in the other 94 studies NK cells were used in combination with chemotherapy, pro‐inflammatory cytokines (mainly IL‐2 or IL‐15), or antibodies (targeting tumor antigens or immune checkpoint inhibitors). In nearly two‐thirds of the studies (114 out of 177), NK cells were not genetically modified (autologous or allogeneic), while 53 out of 177 trials used CAR‐NK cells and only 10 NK‐92 cell line. For 49 studies, efficacy results could be found on the Beacon database (data.beacon‐intelligence.com), and they are expressed as overall response rate (ORR) in Figure . Table summarizes interventions for each clinical trial. The number of clinical trials reporting efficacy outcomes remains limited for each tumor indication. In addition, substantial variability in clinical responses was observed across studies, potentially reflecting differences in prior lines of therapy received before NK cell infusion. As a result, no single tumor indication or specific intervention could so far be clearly identified as demonstrating superior clinical benefit. Clinical efficacy of NK cell infusions in patients with solid tumors. Clinical activity was assessed using the overall response rate (ORR), defined as the proportion of patients with complete or partial responses among all evaluable patients. Each bubble represents a clinical trial, and the size of the bubbles represents the number of evaluable patients. 【Clustering of Tumor Indications Based on Transcriptomic Data to Identify Functional NK Cell‐Enriched Tumors】 To further explore the solid tumor conditions presenting the highest chances to respond to NK cell therapies, we analyzed the various RNA‐seq datasets available from the TCGA database of more than 10,000 tumor samples for 31 different types of cancer to assess NK‐cell infiltration, NK‐cell fitness, and microenvironmental factors (Table ). Informative genes significantly overexpressed by NK cells (CHST12, KLRF1, GNLY, KLRC1, IL18RAP, KLRD1, IL2RB, CST7, XCL2, CD244, PRF1, NCR1, KIR2DL1, KIR2DL3, CD160, and KLRC3) [] were selected to infer the abundance of these cell types in the tumor samples analyzed. By heatmap analysis, a group of tumors characterized by high NK cell infiltration emerged (Figure ). Then, the expression of the activating NK cell receptors NKp30, CD16, KLRF1, CD160, DNAM1, KLRK1, CD244 and of the cytotoxic molecules perforin and granzyme B was used to identify tumors containing NK cells with the highest cytotoxic potential [] (Figure ). Within the tumor microenvironment, the balance of ligands for activating and inhibitory NK cell receptors, along with immunosuppressive soluble factors such as TGF‐β, PGE2, and IL‐10, shapes NK cell function [, ]. Analysis of the expression of these molecules (Figure ) delineated a subset of tumors exhibiting microenvironmental features favorable to NK cell‐mediated lysis, including kidney renal clear cell carcinoma, lung adenocarcinoma and squamous cell carcinoma, mesothelioma, pancreatic adenocarcinoma, cervical and endocervical cancers, head and neck squamous cell carcinoma, bladder urothelial carcinoma and sarcoma. Notably, lung, kidney, cervical and endocervical cancers and mesothelioma also displayed substantial infiltration by functional NK cells (Figure ). Collectively, these findings suggest that kidney renal cell carcinoma, lung adenocarcinoma and squamous cell carcinoma, cervical and endocervical cancers, and mesothelioma may be promising candidates for NK cell‐based therapies (Figure ). However, it should be noted that these conclusions are based on bulk transcriptomic analyses, which are inherently limited by the heterogeneity of tumor samples, in terms of immune cell infiltration, tumor stage, and prior treatments. Solid tumor indication selection by clustering from transcriptomic dataset analysis. By heatmap analysis of RNA‐seq datasets from the TCGA database across 31 solid tumor types, assessing the expression of genes indicating NK cell presence and fitness, activating NK cell receptor ligands, and NK cell inhibitory molecules, a selection of tumor indications most likely to respond to NK cell‐based therapies was proposed. First, the expression of genes significantly overexpressed by NK cells (CHST12, KLRF1, GNLY, KLRC1, IL18RAP, KLRD1, IL2RB, CST7, XCL2, CD244, PRF1, NCR1, KIR2DL1, KIR2DL3, CD160, KLRC3) [] and the expression of the activating NK cell receptors NKp30, CD16, KLRF1, CD160, DNAM1, KLRK1, CD244 and of the cytotoxic molecules perforin and granzyme B were analyzed to identify the presence of cytotoxic NK cells (left). Then, the presence of positive or negative NK cell regulators in the tumor microenvironment (ligands for activating and inhibitory NK cell receptors and soluble factors such as TGF‐β, PGE2, and IL‐10) was evaluated to detect tumors exhibiting microenvironmental features favorable to NK cell‐mediated lysis (right). Tumor types identified by both analyses are highlighted in bold. a, adenocarcinoma; cc, cell carcinoma; c, carcinoma. Indications characterized by high intratumoral NK‐cell infiltration and fitness may be particularly suitable for therapeutic strategies leveraging endogenous NK cells in vivo (e.g., antibodies or immunocytokines). In contrast, indications displaying favorable expression patterns of activating ligands together with limited inhibitory signals may represent promising settings for adoptive NK‐cell infusion approaches, whether genetically engineered or not. The expression of the chemokines that mediate the recruitment of NK cells in tumors, i.e., CX3CL1, RARRES2 (Chemerin), CCL5, CXCL9, CXCL10 and CXCL14 [, , , , ], was also analyzed, but no major differences were observed among the different tumor indications (Figure ). The wide majority of solid tumors seem to express NK cell‐attracting chemokines; nevertheless, NK cells represent a minor population of tumor‐infiltrating lymphocytes. Potential explanations include failure to reach the required threshold, physical barriers limiting NK cell infiltration, or a lack of growth factors required to sustain NK cell viability [, , ]. 【Conclusions】 Collectively, these data provide an integrative multiomic framework for the rational prioritization of solid tumor indications for NK cell‐based therapies. By systematically combining immunohistochemical evidence of NK cell infiltration, clinical trial efficacy data, transcriptomic profiling, and tumor microenvironment characterization, we identify kidney renal clear cell carcinoma, lung adenocarcinoma and squamous cell carcinoma, cervical and endocervical cancers, and mesothelioma as the tumor types presenting the most favorable biological landscape for NK cell‐mediated antitumor activity. These indications are characterized by meaningful NK cell infiltration and a tumor microenvironment with a favorable balance of activating ligands and limited immunosuppressive signaling. The clinical trial landscape analysis highlights the current heterogeneity and fragmentation of the NK cell therapy field in solid tumors, with no single intervention or indication yet demonstrating unequivocal superiority. These findings likely reflect the heterogeneity of patients' prior treatment histories and the frequently limited number of evaluable patients in each study. This underscores the urgent need for biomarker‐driven patient selection and standardized endpoints to enable meaningful cross‐study comparisons. Our transcriptomic analyses begin to address this gap by proposing objective, data‐driven criteria for indication selection that extend beyond empirical observation. Taken together, these findings have direct translational implications. They provide a biologically grounded rationale for focusing NK cell clinical development on a defined set of solid tumor indications, thereby improving the probability of clinical success and accelerating the path toward regulatory approval. As NK cell engineering continues to advance—with increasingly sophisticated CAR constructs, cytokine armoring, and strategies to overcome microenvironmental immunosuppression—the framework presented here offers a roadmap for aligning these innovations with the solid tumor contexts in which they are most likely to be effective. Prospective validation of these findings in dedicated biomarker‐enriched clinical trials will be essential to translate this multimodal evidence into improved patient outcomes. 【Author Contributions】 L.C., V.P., T.A.F. and E.V. conceived the presented data. L.C. performed immunohistochemistry data analysis and reviewed clinical trial results. L.C. and Q.B. analyzed and interpreted RNA‐seq data. L.C., V.P., Q.B, T.A.F., and E.V. discussed the results and reviewed the manuscript. 【Conflicts of Interest】 The authors declare no conflicts of interest. 【Supporting information】 Supporting File 1: eji70283‐sup‐0001‐SuppMat.pdf. Supporting File 2: eji70283‐sup‐0002‐Table S1.xlsx. Supporting File 3: eji70283‐sup‐0003‐Table S2.xlsx.