Clinical Readiness of Additively Manufactured Dental Ceramics for Crowns, Veneers, and Partial-Coverage Restorations: A Scoping Review and Evidence Map

Abstract

This scoping review mapped the clinical readiness of directly additively manufactured (AM) dental ceramics for single-unit definitive restorations (crowns, veneers, and partial-coverage restorations) using a predefined review-specific five-tier readiness framework (R1–R5) designed to organize evidence maturity from restoration-relevant foundational studies to comparative clinical evidence. MEDLINE (PubMed), Scopus, Web of Science Core Collection, EBSCO (Dentistry and Oral Sciences Sources), and ClinicalTrials.gov were searched from inception to February 2026, with citation tracking. Thirty-five sources were included: 31 in vitro studies and 4 clinical studies. Evidence clustered in preclinical tiers, with most studies classified as restoration-level in vitro investigations (R2, 22/35) or foundational specimen-level studies explicitly linked to restorative performance (R1, 9/35); only one feasibility study reached R3 (1/35), three studies provided comparative clinical evidence (R4, 3/35), and no R5-level evidence was identified. The additively manufactured definitive restorations evaluated were zirconia-based. Most restoration-level studies addressed zirconia crowns (18/35), with fewer studies focusing on veneers/laminates (5/35) and occlusal veneers/tabletops (2/35). Across AM routes (most commonly vat photopolymerization ceramic workflows and nanoparticle jetting) outcomes focused on fit/adaptation, manufacturing accuracy, mechanical performance, and aging simulations; clinical studies reported short- to mid-term performance using standardized evaluation criteria. Overall, the evidence suggests technical feasibility and increasing restoration-level evaluation under controlled conditions, but clinical applicability remains preliminary because higher-readiness clinical evidence is still limited. Future work should prioritize standardized reporting, clinically relevant aging/fatigue paradigms, and longer-term comparative clinical studies.

1. Introduction

The expansion of digital dentistry has increased interest in additive manufacturing (AM) as an alternative to subtractive fabrication for indirect restorations [1,2,3]. Subtractive methods, especially computer-aided design/computer-aided manufacturing (CAD/CAM) milling, remain the reference workflow for precision and material reliability [4]. However, AM offers important advantages, including greater geometric freedom, lower material waste, and the ability to produce structures that are difficult to achieve by milling [5]. These advantages have encouraged investigation of AM across multiple dental applications where material efficiency, restoration design flexibility, and workflow customization are increasingly important [6,7]. Nevertheless, the direct translation of AM to high-performance dental ceramics remains technically demanding and depends on careful control of material properties, processing parameters, and post-processing steps [8].

AM, commonly referred to as 3D printing, includes several process families that build objects layer by layer from a digital design [9,10,11]. It can be applied to metals, polymers, and ceramics, and is often associated with design flexibility, surface detail, mass customization, and reduced material waste compared with subtractive fabrication. In ceramic AM, the printed output is typically a ceramic-loaded green body that requires debinding and sintering, and in some glass-ceramic systems additional crystallization, before clinically relevant density and mechanical performance can be achieved [2,12] (Figure 1). As a result, final restoration quality depends not only on printer resolution, but also on slurry characteristics, particle packing, shrinkage control, thermal processing, and defect formation [13,14]. These factors directly affect fit, reliability, and fracture behavior, all of which are critical for brittle restorative materials [15].

Hence, direct AM of ceramic and glass-ceramic definitive restorations (rather than printed polymer patterns intended for subsequent pressing or casting) poses additional manufacturing constraints that directly affect fit, densification, defect formation, and brittle mechanical behavior [10,16]. For this reason, direct ceramic AM requires evaluation beyond isolated proof-of-concept demonstrations before routine clinical use can be considered [8,17].

The interest in directly additively manufactured ceramics is driven partly by the limitations of subtractive milling, including material waste, bur-access restrictions, and the geometric constraints imposed by prefabricated blanks [18]. Subtractive processes are further constrained by the dimensions of prefabricated blanks, the introduction of microcracks from cutting tools, and the inability to fabricate controlled gradient porosities or geometries that are achievable through layer-by-layer deposition [1,19,20]. Furthermore, AM may in principle enable graded internal architectures and localized optical or material variation; however, in definitive dental ceramics, these possibilities remain largely experimental and are not yet broadly established in routine clinical workflows [6,19].

Contemporary prosthodontics relies heavily on all-ceramic materials, such as zirconia, to satisfy the increasing esthetic demands and functional requirements for crowns, veneers, and partial-coverage restorations [20,21]. For implant-supported indications, restoration performance should also be interpreted within a broader biomechanical context, because masticatory load transfer at the implant–abutment complex remains relevant to long-term restorative behavior [22]. However, the integration of AM for definitive ceramic restorations remains at an early stage. Much of the available literature still focuses on in vitro evaluation of fit, trueness, fracture-related outcomes, or material behavior, whereas clinically relevant translation depends on a broader understanding of evidence maturity across indications, manufacturing routes, and post-processing pathways [10,23].

The primary aim of this scoping review is to map the published evidence on directly additively manufactured ceramic and glass-ceramic restorations for crowns, veneers, and partial-coverage restorations. Specifically, the review sought to characterize the available evidence by indication, material class, and AM route and to organize that evidence using a predefined five-tier clinical readiness framework (R1–R5); and to identify translational gaps between preclinical evaluation and clinical use. In line with the scoping review approach, the purpose was to describe the extent of the evidence base rather than to make pooled comparative effectiveness claims.

2. Materials and Methods

2.1. Protocol and Reporting Framework

This review was conducted as a scoping review in accordance with Joanna Briggs Institute (JBI) guidance for scoping reviews and is reported following the PRISMA Extension for Scoping Reviews (PRISMA-ScR) checklist [24,25] (see Supplementary File S1). A prespecified protocol guided eligibility criteria, searching, study selection, data charting, and evidence mapping. No formal critical appraisal or risk-of-bias assessment was undertaken, as the purpose of this review was to map the extent, characteristics, and clinical readiness of the available evidence rather than to estimate certainty of effect. The PICO question was the following: In patients receiving tooth- or implant-supported single-unit definitive restorations (crowns, veneers, and partial-coverage restorations), and in laboratory models simulating these restorations, does direct AM of ceramic/glass-ceramic definitive restorations, compared with conventional fabrication workflows (e.g., CAD/CAM milling or other conventional ceramic routes as defined by study authors), achieve comparable clinical performance/complications and restoration-relevant laboratory outcomes (including fit/trueness/adaptation, fracture/fatigue behavior, aging/wear performance, and related evaluation metrics)?

P (Population/Problem): Patients receiving tooth- or implant-supported single-unit definitive restorations (crowns, veneers, partial-coverage restorations), and laboratory models simulating these restorations (extracted teeth/typodonts/dies/models).

I (Intervention): Definitive restorations directly fabricated by ceramic AM (printed ceramic/glass-ceramic green body followed by debinding/sintering ± crystallization), including NPJ/material jetting and vat photopolymerization slurry printing (LCM/DLP/SLA) and other eligible ceramic AM routes.

C (Comparator): Conventional fabrication workflows for the same indications (e.g., CAD/CAM milling and/or other conventional ceramic manufacturing routes as defined by study authors), or no comparator where designs are single-arm.

O (Outcomes): Clinical performance and complications (when available), and restoration-relevant laboratory outcomes including fit/trueness/adaptation, mechanical performance (fracture/fatigue), aging/wear behavior, and other evaluation metrics reported for eligible indications.

2.2. Eligibility Criteria

This review includes clinical studies of human participants receiving tooth- or implant-supported, single-unit definitive ceramic restorations and laboratory studies using extracted teeth, typodonts, dies, or models. Eligible interventions were definitive restorations directly produced via ceramic AM within prosthodontic/restorative workflows or laboratory simulations of those workflows. Indications included crowns, veneers (including ultrathin/partial-prep), and partial-coverage restorations (inlays, onlays, overlays, occlusal veneers). Eligible materials comprised zirconia and dental glass-ceramics (including lithium disilicate and lithium-silicate-derived variants), and eligible AM routes included nanoparticle/material jetting and vat photopolymerization of ceramic slurries (e.g., LCM/DLP/SLA slurry printing), as well as other ceramic AM processes when the final restoration was ceramic. Included study designs spanned clinical interventional and observational studies (including case series/reports) and laboratory comparative, validation, fatigue/fracture/wear/aging, and fit/trueness studies; secondary reviews were used for background only. The review was limited to peer-reviewed English-language studies published up to 1 February 2026, and excluded indirect workflows (printed patterns for pressing/casting), non-ceramic “permanent crown” polymers/composites, non-restorative applications, and coupon/disk-only studies lacking explicit restoration relevance. The inclusion criteria are presented in

Table 1. Inclusion criteria.

Domain	Inclusion Criteria
Publication type	Peer-reviewed journal primary research articles (clinical or in vitro)
Language	English
Timeframe	From database inception to final search date
Concept (index intervention)	Definitive restorations directly fabricated by AM as a ceramic/glass-ceramic body and densified via debinding/sintering ± crystallization
Eligible indications	Single-unit crowns; veneers; partial-coverage restorations (inlays, onlays, overlays, occlusal veneers) as defined by authors
Eligible material classes	Zirconia; dental glass-ceramics used for eligible indications (e.g., lithium disilicate, lithium silicate-derived glass-ceramics)
Eligible AM routes	Material jetting/nanoparticle jetting; vat photopolymerization of ceramic slurries (LCM; DLP/SLA slurry); other ceramic AM routes if the final definitive restoration is ceramic
Clinical evidence (if applicable)	Human participants receiving eligible ceramic restorations (tooth-supported or implant-based)
Laboratory evidence (if applicable)	Restoration-level specimens fabricated for extracted teeth, typodonts, dies, or models, evaluated with restoration-relevant outcomes (fit/trueness; mechanical/fatigue; aging/wear; defect characterization; failure modes)
Coupon/disk-only laboratory evidence	Include only if coupon/disk tests explicitly link processing to properties relevant to eligible indications

Abbreviations: AM, additive manufacturing; LCM, lithography-based ceramic manufacturing; DLP, digital light processing; SLA, stereolithography.

.

2.3. Search Strategy

The following databases were searched: MEDLINE (PubMed), Scopus, Web of Science Core Collection, Dentistry & Oral Sciences Source, and ClinicalTrials.gov. Backward and forward citation tracking of included studies and key reviews was performed to identify additional eligible records. A comprehensive search strategy combined terms for (1) AM/3D printing routes, (2) ceramic material classes, and (3) prosthodontic indications. Controlled vocabulary (e.g., MeSH) and free-text terms were used as appropriate and adapted for each database. Full search strategies for all databases are provided in

Table 2. Search strategies across databases.

Database	Interface/Fields Searched	Search String (Copy/Paste)	Limits/Filters
MEDLINE (PubMed)	PubMed; MeSH + Title/Abstract ([tiab])	(“Printing, Three-Dimensional” [Mesh] OR 3D print*[tiab] OR three-dimensional print*[tiab] OR additive manufactur*[tiab] OR lithography-based ceramic manufacturing[tiab] OR LCM[tiab] OR nanoparticle jetting[tiab] OR NPJ[tiab] OR ceramic stereolithograph*[tiab] OR vat photopolymer*[tiab] OR DLP[tiab] OR “digital light processing”[tiab] OR SLA[tiab] OR material jetting[tiab] OR binder jet*[tiab]) AND (zirconia[tiab] OR Y-TZP[tiab] OR yttria-stabilized zirconia[tiab] OR “lithium disilicate”[tiab] OR “lithium silicate”[tiab] OR glass-ceramic*[tiab]) AND (crown*[tiab] OR veneer*[tiab] OR “laminate veneer”[tiab] OR inlay*[tiab] OR onlay*[tiab] OR overlay*[tiab] OR “occlusal veneer”[tiab] OR “partial coverage”[tiab])	Language: English; Date range: inception–1 February 2026
Web of Science Core Collection	Advanced Search; Topic (TS=)	TS=(((“3D print*” OR “three-dimensional print*” OR additive manufactur* OR “rapid prototyp*” OR “lithography-based ceramic manufacturing” OR (LCM NEAR/3 ceramic*) OR “nanoparticle jetting” OR (NPJ NEAR/3 ceramic*) OR “ceramic stereolithograph*” OR “vat photopolymer*” OR (DLP NEAR/3 (ceramic* OR slurry)) OR (SLA NEAR/3 (ceramic* OR slurry)) OR “digital light processing” OR “material jetting” OR “binder jet*”) AND (zirconia OR “yttria-stabilized zirconia” OR Y-TZP OR “lithium disilicate” OR “lithium silicate” OR glass-ceramic*) AND (crown* OR veneer* OR “laminate veneer” OR inlay* OR onlay* OR overlay* OR “occlusal veneer” OR “partial coverage”)))	Language: English; Timespan: inception–2026; Final search date recorded as 1 February 2026
Scopus	Document search; TITLE-ABS-KEY ()	TITLE-ABS-KEY(((3d W/1 print* OR “three-dimensional” W/1 print* OR “additive manufactur*” OR “rapid prototyp*” OR “lithography-based ceramic manufacturing” OR (lcm W/3 ceramic*) OR “nanoparticle jetting” OR (npj W/3 ceramic*) OR “material jetting” OR “binder jet*” OR “ceramic stereolithograph*” OR “vat photopolymer*” OR “ceramic slurry” OR (dlp W/3 (ceramic* OR slurry)) OR (sla W/3 (ceramic* OR slurry)) OR “digital light processing”) AND (zirconia OR “yttria-stabilized zirconia” OR y-tzp OR “lithium disilicate” OR “lithium silicate” OR glass-ceramic*) AND (crown* OR veneer* OR “laminate veneer” OR inlay* OR onlay* OR overlay* OR “occlusal veneer” OR “partial coverage”))))	Language: English; Date range: inception–1 February 2026 (via Scopus year/date filters)
Dentistry & Oral Sciences Source (EBSCOhost)—additional database (protocol amendment)	EBSCOhost Advanced Search; two rows (TI and AB) combined with OR; paste the same string into TI and again into AB	(((“3D print*” OR “three-dimensional print*” OR “additive manufactur*” OR “lithography-based ceramic manufacturing” OR “nanoparticle jetting” OR NPJ OR “material jetting” OR “binder jet*” OR LCM OR DLP OR SLA OR “digital light processing”) N5 (zirconia OR “yttria-stabilized zirconia” OR Y-TZP OR “lithium disilicate” OR “lithium silicate” OR glass-ceramic*)) AND (crown* OR veneer* OR “laminate veneer” OR inlay* OR onlay* OR overlay* OR “occlusal veneer” OR “partial coverage”))	Language: English; Peer-reviewed; Journals; Date range: inception–1 February 2026

Abbreviations: MeSH, medical subject headings; tiab, title/abstract; TS, topic search; TITLE-ABS-KEY, title/abstract/keywords; TI, title; AB, abstract; LCM, lithography-based ceramic manufacturing; NPJ, nanoparticle jetting; DLP, digital light processing; SLA, stereolithography; Y-TZP, yttria-stabilized tetragonal zirconia polycrystal.

.

In addition, a supplementary search was conducted in Elicit Pro (Ought, Oakland, CA, USA) using the same core concept blocks (AM/3D printing routes, ceramic material classes, and prosthodontic indications); candidate records returned by the tool (including citation recommendations) were exported, de-duplicated against database records, and screened using the same eligibility criteria as all other records (

Table 3. Elicit search strategy.

Item	Description
Tool/coverage	Elicit semantic search across Semantic Scholar and OpenAlex corpora (supplementary method).
Final search date (recorded)	1 February 2026.
Elicit query	What are the clinical outcomes and failure modes of directly 3D-printed zirconia or lithium disilicate/lithium-silicate-derived glass-ceramics single-unit restorations compared to conventionally manufactured zirconia crowns?
Retrieval	The search returned 1000 total results from Elicit. The 1000 most relevant were retrieved for screening.
Abstract screening	Strict yes/no screening criteria were applied. Papers failing any strict criterion were automatically excluded.
Data extraction approach	LLM-assisted column-based extraction using structured prompts (e.g., study type; restoration specifications; outcomes and measurement methods; failure modes and analysis; manufacturing route and parameters; post-processing; surface treatments; defect signatures; readiness tier R1–R5).

Abbreviations: LLM, large language model; R1–R5, readiness tiers used in the review framework.

).

2.4. Study Selection

Records were imported into ENDNOTE X9 (Clarivate, Philadelphia, PA, USA) for de-duplication. Four independent reviewers screened titles/abstracts against the eligibility criteria. Full texts were retrieved for potentially eligible records and independently assessed by four reviewers. Disagreements were resolved by discussion; a fifth reviewer adjudicated when required.

2.5. Data Charting

A standardized data charting form was developed and piloted on an initial subset of included studies and refined iteratively. Data charting was performed by reviewers using Elicit Pro (Ought, Oakland, CA, USA) with any automatically pre-populated fields verified against the full text to ensure accuracy. Specifically, all AI-assisted outputs were independently checked by two human reviewers against the source PDFs and extraction sheets; no eligibility decision, study characteristic, numerical value, or citation was accepted without manual confirmation.

Core charting items included:

• Bibliometrics (publication year; country/region; setting; funding and conflicts of interest as reported);
• Indication (crown; veneer; partial-coverage subtype);
• Material class and details (e.g., zirconia grade as reported; glass-ceramic type; translucency/shade system when provided);
• AM route and key parameters (e.g., build orientation; layer thickness; support strategy; green body handling), as reported;
• Post-processing (debinding/sintering/crystallization schedules; shrinkage compensation approach, when described);
• Finishing/surface treatment (polishing/glazing; internal surface conditioning; cementation protocol, if applicable);
• Comparator(s) (e.g., milled/pressed ceramics; alternative manufacturing routes);
• Outcome domains (fit/adaptation; trueness/precision; mechanical outcomes including fracture/fatigue; aging protocols; wear/antagonist wear; and clinical outcomes);
• Defect signatures/characterization (e.g., porosity/density via micro-CT; microcracks via SEM; phase composition via XRD; surface roughness/topography);
• Failure modes and failure analysis (e.g., fracture/chipping location; debonding; marginal breakdown; fractography when reported);
• Assigned clinical readiness level (R1–R5).

2.6. Clinical Readiness Framework (R1–R5)

Each included source was assigned a clinical readiness level based on the most clinically proximal evidence presented, using a five-tier framework ranging from foundational restoration-relevant material evidence (R1) to longer-term or pragmatic clinical evidence (R5):

R1: Foundational material evidence (coupons/disks) with explicit restoration-relevant linkage;

R2: Restoration in vitro (crowns/veneers/onlays tested on dies/teeth/models);

R3: Human feasibility (case report/series; short follow-up);

R4: Comparative clinical evidence (controlled studies or randomized trials);

R5: Longer-term, pragmatic or multi-center clinical evidence.

2.7. Synthesis and Presentation of Results

Results are presented using: (i) a PRISMA-ScR flow diagram and descriptive summary of included sources; (ii) an evidence map structured as material class × AM route × indication, annotated by readiness level and volume of evidence; (iii) a narrative synthesis by indication (crowns, veneers, partial-coverage) highlighting readiness drivers and limitations; (iv) a process–defect–failure table linking manufacturing and post-processing features to defect signatures and reported failure modes; and (v) a gap analysis leading to a prioritized research agenda (e.g., reporting standardization for post-processing, clinically meaningful aging/fatigue protocols, longer follow-up, and comparator selection). Quantitative synthesis (meta-analysis) was not planned due to expected heterogeneity. ChatGPT 5.2 (OpenAI) and Elicit (Ought) were used to draft standardized, non-decisional prose, to generate schematic or illustrative figures, and to structure extraction templates; they were not used to make eligibility determinations, populate data fields, derive numerical values, or finalize study characteristics without human verification. A predefined validation protocol governed every AI-assisted artifact and required line-by-line human editing by two reviewers, cross-checking of all statements, attributes, and numbers against the extraction sheets and the source PDFs, rejection of any suggested citation not retrievable from the registered search sources, and archiving of prompts and outputs with timestamps; final acceptance of AI-assisted text or graphics required explicit agreement by at least two human reviewers.

3. Results

3.1. Study Selection

A total of 1230 records were identified through database searching and other sources. After removal of duplicates, 916 records were screened by title/abstract. Overall, 147 full-text reports were assessed for eligibility. In total, 112 full-text reports were excluded (with reasons). A total of 35 sources were included in the scoping review (Figure 2).

3.2. Characteristics of Included Sources

Overall, the included evidence base was dominated by zirconia-focused studies and clustered in the preclinical readiness tiers, with crowns representing the most frequently investigated indication and only limited progression into clinical comparative research. Of the 35 studies included, most sources were in vitro (31/35), and 4 were clinical (

Table 4. Characteristics of included studies.

Study	Study Type	AM Technology	Comparator	Restoration Type	Tooth Location	n Per Group	Aging Protocol	Readiness Tier
Fangyuan et al., 2025 [26]	RCT	NPJ	CAD/CAM milling	Posterior crown	Posterior	24	None	R4
Refaie et al., 2023 [30]	In vitro	DLP (CeraFab7500)	Milling (DWX520)	Premolar crown	Upper premolar	15	1.2 M mechanical cycles	R2
Hassan et al., 2023 [31]	In vitro	DLP (CeraFab S65)	Milling (CORiTEC 150i)	Molar crown	Mandibular first molar	12	5000 thermocycles	R2
Refaie et al., 2023 [32]	In vitro	DLP (CeraFab7500)	Milling (DWX520)	Premolar crown	Upper premolar	10	None	R2
Zhu et al., 2023 [33]	In vitro	NPJ (Carmel 1400)	Milling (AMD500)	Molar crown	Mandibular first molar	16	None	R2
Ioannidis et al., 2021 [34]	In vitro	LCM (CeraFab 7500)	Milling + heatpress (LS2)	Occlusal veneer	Molar	20	None	R2
Camargo et al., 2022 [35]	In vitro	NPJ (Carmel 1400)	Chairside + lab milling	Molar crown	Maxillary molar	10	None	R2
Zandinejad et al., 2021 [20]	In vitro	SLA (CeraMaker 900)	None (AM vs. AM)	Implant crown	Maxillary second premolar	10	None	R2
Abualsaud et al., 2022 [36]	In vitro	SLA (CERAMAKER C900)	Milling (PrograMill PM7)	Molar crown	Mandibular first molar	10	None	R2
Ioannidis et al., 2020 [37]	In vitro	LCM (CeraFab 7500)	Milling + heatpress (LS2)	Occlusal veneer	Molar	20	1.2 M cycles + thermocycling	R2
Handermann et al., 2024 [38]	In vitro	LCM (CeraFab 7500)	Milling (Cercon brain expert)	Anterior partial crown	Anterior	15	None	R2
Kalman et al., 2024 [39]	In vitro	LCM (CeraFab 7500)	Milling (Glidewell)	Anterior crown + veneer	Central incisors	6	None	R2
Revilla-León et al., 2020 [40]	In vitro	SLA (CERAMAKER 900)	Milling (Straumann CARES)	Implant crown	First premolar	10	None	R2
Teegen et al., 2024 [41]	In vitro	DIW	Milling (CADfirst)	Disk specimens	Not applicable	20	None	R1
Branco et al., 2020 [42]	In vitro	Robocasting	Milling (Zirkonzahn M5)	Disk specimens	Molar cusps as antagonist	8	Chewing simulation	R2
Cho et al., 2025 [43]	In vitro	DLP + SLA (multiple)	Milling (5-axis)	Crown	Not specified	10	None	R2
Zenthöfer et al., 2024a [44]	In vitro	LCM + DLP (CeraFab 7500, Zipro-D)	Milling (PrograMill PM7)	Occlusal veneer	Molar	8	Thermocycling + 1.2 M chewing cycles	R2
Lee et al., 2024 [45]	In vitro	SLA + DLP	Milling (Datron D5)	Molar crown	Maxillary first molar	10	None	R2
Hetzler et al., 2025 [46]	In vitro	DLP (ZiproD)	Milling (PrograMill PM7, Cercon)	Disk specimens	Posterior	≥23	None	R1
Elsayed et al., 2025 [47]	In vitro	DLP	Milling (DWX52 DI)	Premolar crown	Premolar	11	5000 thermocycles + 75,000 mechanical cycles	R2
Moon et al., 2022 [48]	In vitro	DLP	Milling (3 systems)	Crown	Not specified	10	None	R1
Yu et al., 2024 [28]	Retrospective clinical	3D gel deposition	Pressed + milled (LS2)	Veneer	Various	40–45	None (clinical)	R4
Jung et al., 2023 [49]	In vitro	DLP (custom)	CAD/CAM (reference)	Crown	Not specified	Not specified	None	R1
Sun et al., 2022 [50]	In vitro	3D gel deposition	Milling (zirconia + LS2)	Molar crown	Mandibular first molar	10	None	R2
Cui et al., 2020 [27]	Clinical trial	3D gel deposition	Milling (Zenotec)	Premolar crown	Premolar	27	None (clinical)	R4
Mohammed et al., 2024 [51]	In vitro (material characterization)	DLP (Zipro)	Conventional machining	Disk specimens	Not applicable	Not specified	None	R1
Sasany et al., 2025 [52]	In vitro	LCM (CeraFab S65)	Milling (CEREC Primemill)	Laminate veneer	Maxillary central incisor	20	5 M chewing cycles + thermocycling	R2
Zhao et al., 2025 [53]	In vitro	DLP (custom)	Milling (ARUM 5X-400)	Ultrathin veneer	Not specified	10–18	LTD autoclave aging	R2
Zhang et al., 2023 [54]	In vitro	DLP (AUTOCERAM)	Dry-pressed zirconia	Crown	Not specified	Not specified	None	R2
Noh et al., 2024 [55]	In vitro	DLP	Milling (K5+)	Laminate	Central maxillary incisor	10	None	R2
Kao et al., 2023 [29]	Clinical pilot	SLM	None (single-arm)	Posterior crown	Premolars and molars	15	None (clinical)	R3
Hajjaj et al., 2024 [56]	In vitro	FDM	Milling (Ceramill ZI)	Bar + disk specimens	Not applicable	15 bars, 10 disks	5000 thermocycles	R1
Lee et al., 2022 [57]	In vitro	SLA (CERAMAKER 900)	Different build directions	Molar crown	Maxillary first molar	10	None	R2
Zenthöfer et al., 2024b [58]	In vitro	DLP (ZiproD, CeraFab7500)	None (printer comparison)	Disk specimens	Not applicable	20	None	R1
Bömicke et al., 2024 [59]	In vitro	LCM (Lithoz)	Milling (IPS e.max ZirCAD)	Disk specimens	Not applicable	14	7500 thermocycles + 30 d water	R1

Abbreviations: AM, additive manufacturing; RCT, randomized controlled trial; NPJ, nanoparticle jetting; DLP, digital light processing; LCM, lithography-based ceramic manufacturing; SLA, stereolithography; DIW, direct ink writing; SLM, selective laser melting; FDM, fused deposition modeling; LTD, low-temperature degradation; CAD/CAM, computer-aided design/computer-aided manufacturing; LS2, lithium disilicate; R1–R4, readiness tiers represented in this table; M, million.

). The clinical evidence comprised one randomized controlled trial [26], one self-controlled clinical trial [27], one retrospective comparative study [28], and one single-arm short-term pilot study [29]. Follow-up windows ranged from 24 weeks [29] to 12 months [26], including a mean follow-up of 12.2 ± 3.5 months [27] and 35.0 ± 14.7 months [28]. Outcome assessment methods were heterogeneous across studies, particularly for marginal and internal fit, manufacturing accuracy, mechanical testing, and aging simulation, which limits direct comparison across AM routes, materials, and restorative indications.

Indications/restoration types: Restoration-level evidence addressed crowns (18/35), veneers/laminates (5/35), occlusal veneers/tabletops (2/35), and partial-coverage crowns (1/35). Foundational studies (9/35) evaluated coupon/specimen-level properties with explicit restorative relevance. Representative restoration-level evaluations included zirconia crown fit/adaptation and margin quality [32,33], occlusal veneer performance [34], and a partial-coverage crown thickness scenario for minimally invasive therapy [38].

Material classes represented: The additively manufactured definitive restorations evaluated in the included sources were zirconia-based. Glass-ceramics (e.g., lithium disilicate) and other ceramics appeared primarily as comparators in some restoration-level studies (e.g., occlusal veneers and laminate veneer comparisons) rather than as directly additively manufactured definitive restorations [28,37,52].

AM routes represented: The included literature encompassed multiple AM routes (Figure 3) for zirconia, including vat photopolymerization ceramic workflows (e.g., DLP/SLA/LCM) used for crown/veneer fabrication and fit/accuracy assessment [33,40,57], nanoparticle jetting for clinical crown fabrication [26], wet/gel deposition for clinical crown and veneer applications [27,28,50], as well as other approaches evaluated primarily at specimen level (e.g., direct ink writing and fused deposition modeling) [41,56].

Comparators. Most comparative studies included CAD/CAM-milled zirconia as the primary comparator [26,32]. Additional comparators included conventionally processed non-zirconia ceramics (e.g., lithium disilicate in veneer/occlusal veneer contexts) and alternative zirconia systems/processing variants [20,37,52].

3.3. Clinical Readiness Distribution

Using the project’s five-tier readiness framework (R1–R5) (Figure 3), the included evidence concentrated in the preclinical tiers:

R1 (foundational coupon/specimen-level with explicit restorative linkage): 9/35 (e.g., Bömicke et al., 2024 [59]; Teegen et al. [41], 2023; Zhao et al., 2025 [53]);

R2 (restoration-level in vitro): 22/35 (e.g., Refaie et al., 2023 [30]; Revilla-León et al., 2020 [40]; Zhu et al., 2025 [33]);

R3 (human feasibility; short follow-up): 1/35 (Kao et al., 2023 [29]);

R4 (comparative clinical evidence): 3/35 (Fangyuan et al., 2025 [26]; Cui et al., 2020 [27]; Yu et al., 2024 [28]);

R5 (longer-term/pragmatic/multicenter clinical evidence): 0/35.

Overall, the evidence base was predominantly R1–R2, with limited clinical evaluation (R3–R4) and no R5-level evidence.

3.4. Evidence Map (Material Class × AM Route × Indication)

Evidence clustered most densely in zirconia crowns assessed in vitro (R2), often focusing on fit/adaptation, margin quality, accuracy, and/or fracture resistance [30,33,40]. A smaller cluster addressed veneers/laminates and occlusal veneers with restoration-level outcomes [37,44,55].

Higher-readiness evidence (R3–R4) was limited to zirconia restorations evaluated clinically: an RCT of nanoparticle jetting crowns [26], a self-controlled trial of additively wet-deposited crowns [27], a retrospective comparative veneer study using gel deposition [28], and a short-term pilot study of selectively laser-melted zirconia crowns [29].

3.5. Outcomes Mapped

3.5.1. Fit/Adaptation (Marginal/Internal)

Fit/adaptation was commonly quantified as marginal and/or internal gaps or discrepancies, using approaches such as silicone replica techniques and 3D comparison methods depending on the study [30,40,43,55]. Some studies reported region-specific gap assessments (e.g., marginal, internal, incisal) for laminate/veneer-type restorations [55], while others focused on crown marginal/internal discrepancies and margin quality [32,33] (

Table 5. Marginal and internal fit outcomes for additively manufactured versus milled zirconia restorations: study-level summary and measurement methods.

Study	AM Marginal Gap	Milled Marginal Gap	AM Internal Fit	Milled Internal Fit	Clinical Acceptability	Method
Refaie et al., 2023 [32]	80 ± 30 µm (VMGT); 100 ± 10 µm (SRT)	60 ± 20 µm (VMGT); 60 ± 10 µm (SRT)	Significant differences except axial	Lower values	Yes, both groups	VMGT + SRT
Zhu et al., 2023 [33]	All-crown RMS 21.8 ± 2.8 µm	VITA: 26.5 ± 3.2 µm; UPCERA: 31.7 ± 3.7 µm	Clinically acceptable (p > 0.05)	Clinically acceptable	Yes, all groups	Triple scan
Ioannidis et al., 2021 [34]	95 µm (margin)	65 µm (margin)	184 µm (3D)	120 µm (3D)	Yes	Digital superimposition
Camargo et al., 2022 [35]	113 µm mean cement space	Lab: 102 µm; Chair: 124 µm	64.9% within acceptable range	Lab: 73.1%; Chair: 54.6%	Yes	Micro-CT
Abualsaud et al., 2022 [36]	38.26 ± 4.87 µm	36.68 ± 6.04 µm	46.67 ± 2.80 µm (overall)	44.63 ± 6.24 µm (overall)	Yes (p > 0.05)	Digital superimposition
Revilla-Leon et al., 2020 [40]	AM: 146 µm; SAM: 79.5 µm	CNC: 37.5 µm	AM: 79 µm; SAM: 85 µm	CNC: 73 µm	AM: No; SAM: Yes; CNC: Yes	SRT
Cho et al., 2025 [43]	42.83–81.95 µm	48.45 µm	Not reported	Not reported	Yes, all < 120 µm	Replica technique
Lee et al., 2024 [45]	SLA: 60.60 µm; DLP: 47.33 µm	21.81 µm	SLA: 51.10 µm; DLP: 39.37 µm	23.07 µm	Yes, AM clinically acceptable	Digital superimposition
Elsayed et al., 2025 [47,50]	41–45 µm (before-after aging)	20–25 µm (before-after aging)	Not reported	Not reported	Both within acceptable range	Stereomicroscope
Sun et al., 2022 [50]	SGZ: 56–64 µm	ZZ: comparable; EMX: highest	SGZ: 60.9–75.9 µm	ZZ: 80.3–102.6 µm	Yes, all groups	Direct-view + replica
Jung et al., 2023 [49]	44.4 ± 10.8 µm (RMS)	Not tested	22.8 ± 1.6 µm (RMS)	Not tested	Yes	3D scanning
Noh et al., 2024 [55]	47.1–70.9 µm	40.7–79.7 µm	Not reported	Not reported	Yes, both groups	SRT

Abbreviations: AM, additively manufactured; RMS, root mean square; VMGT, vertical marginal gap thickness; SRT, silicone replica technique; micro-CT, micro-computed tomography; SLA, stereolithography; DLP, digital light processing; CNC, computer numerical control; SAM, subtractive/AM hybrid (as termed by the original study).

).

3.5.2. Manufacturing Accuracy (Trueness/Precision)

Manufacturing accuracy was assessed using 3D deviation-based metrics (commonly RMS-based) and region-specific analyses for crown-like geometries in some studies [33,48], and as internal adaptation uniformity in gel-deposited crowns [50]. Several studies explicitly evaluated the influence of manufacturing route and/or process factors (e.g., build direction or system differences) on accuracy-related outcomes [43,57].

Research gap statement: Accuracy studies differed in reference definitions, alignment strategies, and surface-region reporting, constraining synthesis across AM systems and workflows.

3.5.3. Mechanical Performance (Fracture Resistance; Fatigue)

Restoration-level mechanical performance was frequently reported as fracture resistance/load-to-failure for crowns and related restorations, sometimes incorporating fatigue/cyclic loading protocols [20,38]. Occlusal veneer/tabletop studies assessed load-bearing capacity and fracture resistance under defined loading conditions [37,58]. Crown fracture resistance comparisons between printed and milled monolithic zirconia were also reported following thermocycling in at least one study [31].

At the foundational (R1) level, mechanical outcomes included standardized strength and hardness metrics (e.g., biaxial flexural strength, Vickers hardness) and related material characterization, including process comparisons [41,46,53,56] (

Table 6. Mechanical performance of additively manufactured versus milled ceramic restorations.

Study	AM Fracture Resistance/Strength	Milled Fracture Resistance/Strength	p-Value	Aging Applied	Clinically Acceptable?
Refaie et al., 2023 [30]	1658 N (no aging); 1224 N (after cycling)	1890 N (no aging); 1642 N (after cycling)	p = 0.002	1.2 M cycles	Yes
Hassan et al., 2025 [31]	Comparable (p = 0.26)	Comparable	p = 0.26	5000 TC	Yes
Zandinejad et al., 2021 [20,37]	Monolithic AM: 1243.5 ± 265.5 N	Not tested (AM vs. AM)	p = 0.6	None	Yes
Ioannidis et al., 2020 [37]	3DP: median F_initial 1650 N; F_max 2025 N	CAM: 1250 N; 1500 N	p = 0.0387 (F_initial)	1.2 M cycles + TC	Yes, all groups
Handermann et al., 2024 [38]	3D-printed ZrO2: 1570 ± 661 N	Milled ZrO2: 886 ± 164 N	p = 0.004	None	Yes
Zenthofer et al., 2024a [44]	ZD 0.8 mm on CoCr: 3309 ± 394 N	PM 0.8 mm on CoCr: lower	p < 0.001	TC + 1.2 M cycles	Yes (except 0.4 mm on FRC)
Hetzler et al., 2025 [46]	All AM groups: significantly lower than milled	Highest	p < 0.05	None	Yes
Zhao et al., 2025 [53]	BO: 452.19 ± 70.07 MPa; BM: 400.07 ± 60.10 MPa	Milled: 511.38 ± 54.04 MPa	p < 0.05	LTD autoclave aging	Yes
Zenthofer et al., 2024b [58]	ZiproD: 3Y 768 ± 63 MPa; 5Y 431 ± 51 MPa; CeraFab: 3Y 786 ± 39 MPa; 5Y 468 ± 46 MPa	Not tested	None	None	Yes
Hajjaj et al., 2024 [56]	Bar: 980 ± 132 MPa; Disk: 1023 ± 67 MPa	Bar: 1074 ± 132 MPa; Disk: 1084 ± 67 MPa	p = 0.277	5000 TC	Yes
Teegen et al., 2023 [41]	DIW: 657 ± 51 MPa	Milled: 1022 ± 38 MPa	p < 0.05	None	Yes
Zhang et al., 2023 [54]	DLP: 862.8 ± 104.8 MPa	Dry-pressed: 1065.3 ± 174.0 MPa	p < 0.05	None	Yes

Abbreviations: AM, additively manufactured; TC, thermocycling; LTD, low-temperature degradation; 3DP, 3D-printed; CAM, computer-aided manufacturing; ZrO₂, zirconia; F_initial, initial fracture load; F_max, maximum fracture load; ZD, zirconia dioxide; PM, pressed/milled comparator group; CoCr, cobalt–chromium; FRC, fiber-reinforced composite; BO, build orientation; BM, build mode; 3Y, 3 mol% yttria-stabilized zirconia; 5Y, 5 mol% yttria-stabilized zirconia.

).

Research gap statement: Mechanical testing approaches were heterogeneous (specimen geometry, restoration support conditions, loading configuration, and fatigue/aging inclusion), limiting direct cross-study comparison.

3.5.4. Aging/Wear (Thermocycling/Chewing Simulation; Wear Outcomes)

Aging protocols included thermocycling and cyclic mechanical loading in several restoration-level studies [30,31], alongside storage and/or combined aging approaches in bonding-focused studies [59]. Wear or antagonist-related assessments were reported in both clinical follow-up contexts and in vitro simulations depending on study design [29,42].

Research gap statement: Aging and wear regimens were variably implemented and inconsistently parameterized (e.g., cycles, environment), limiting durability-focused synthesis.

3.5.5. Surface State and Finishing (Roughness/Topography; Polishing/Glazing; Internal Conditioning)

Surface-related outcomes included roughness/topography and surface characteristics, often in relation to manufacturing route and post-processing/finishing [31,43]. Some studies explicitly investigated finishing-related factors (e.g., glaze presence or surface treatment state) alongside functional simulation or material characterization [42,53].

Research gap statement: Finishing and surface-conditioning protocols were variably reported, reducing interpretability across AM workflows.

3.5.6. Bonding/Cementation (Bond Strength; Cement Protocols)

Bonding outcomes were evaluated using resin–zirconia bond strength tests under different conditioning and aging regimens [53,59]. Porcelain-to-zirconia bond strength was also assessed as a function of manufacturing route in at least one study [48].

Research gap statement: Bonding studies differed in conditioning protocols, cement/primer selection, and aging methods, limiting comparisons across studies and settings.

3.5.7. Clinical Outcomes (Survival/Success/Complications; Follow-Up; Evaluation Criteria)

Clinical outcome reporting used structured evaluation criteria (e.g., FDI and modified USPHS/CDA-type frameworks) with follow-up ranging from weeks to months/years. In a 12-month RCT, nanoparticle jetting zirconia single crowns were clinically comparable to CAD/CAM-milled crowns, with reported clinical success rates of 95.8% versus 91.6% and one fracture in each group [26]. In a self-controlled clinical trial, wet-deposited anatomic-contour zirconia crowns showed 100% survival with follow-up after a mean of 12.2 ± 3.5 months [27]. In a retrospective comparative study, 56 patients with 126 veneers were evaluated, with a mean follow-up of 35.0 ± 14.7 months, comparing gel-deposited self-glazed zirconia veneers against lithium disilicate veneer comparators [28]. In a short-term pilot study of additively manufactured zirconia crowns fabricated by selective laser melting, 15 patients were followed for 24 weeks with clinical parameter tracking and reported satisfactory crown grades within the follow-up period [29].

Research gap statement: Clinical evidence remained limited in number and design diversity, with short-to-medium follow-up and no longer-term pragmatic or multicenter evaluations.

3.5.8. Defect Signatures and Failure Modes (Process → Defect → Failure)

Defect-related characterization encompassed microstructural and phase analyses (e.g., SEM and XRD), alongside standardized mechanical testing and, in some studies, failure/fractographic analysis approaches). In restoration-level fatigue/fracture studies, fracture events were commonly the primary failure outcome, with some studies explicitly reporting SEM-based fractographic assessment of fractured specimens [30]. Foundational studies linked process parameters to microstructure-related features (e.g., density/porosity, microstructural appearance, phase composition) and interpreted how these might relate to mechanical response under testing [41,42,53] (

Table 7. Manufacturing and post-processing features linked to defect signatures and failure modes in AM zirconia.

Feature Domain	Process/Post-Processing Feature	Defect Signature (Reported)	Reported Failure Mode (Reported)	Evidence (n; Exemplar Studies)
Manufacturing	Layer-wise AM interfaces (layered build in vat photopolymerization/NPJ)	Porosity aligned with printing layer interfaces	Catastrophic fracture; fracture initiation/propagation associated with layer-interface defects	n = 3; Refaie et al., 2023 [30]; Zenthöfer et al., 2024b [58]; Hetzler et al., 2025 [46]
Manufacturing	Printing orientation (weakest orientation vs. strongest)	Layer-interface alignment/anisotropy	Preferential fracture pathways in the weakest orientation; catastrophic fracture	n = 2; Zenthöfer et al., 2024b [58]; Hetzler et al., 2025 [46]
Manufacturing	Layer boundary microstructure (layer-wise build)	Cracks follow preferential pathways along layer boundaries	Catastrophic fracture propagation along layers	n = 1; Hetzler et al., 2025 [46]
Manufacturing	Vat photopolymerization (DLP)	Delamination/layer separation	Delamination/layer separation during loading	n = 1; Hassan et al., 2025 [31]
Manufacturing	Printed crown geometry (internal surface origin noted in fractography)	Fracture origin at internal (intaglio) surface	Catastrophic fracture (internal origin → outward propagation)	n = 1; Refaie et al., 2023 [30]
Manufacturing	As-printed surface topography (pre-finishing)	Voids + surface irregularities + staircase-like topography (printed vs. milled)	Not linked to a specific failure mode in the synthesis text	n = 3; Cai et al., 2025 [26]; Hassan et al., 2025 [31]; Camargo et al., 2022 [35]
Manufacturing	FDM route (thermoplastic binder extrusion)	Not specified as a distinct defect signature in the synthesis; marked underperformance in mechanical properties	Flexural test failure context: markedly reduced flexural strength vs. milled; “not currently suitable for definitive restorations”	n = 1; Hajjaj et al., 2024 [56]
Post-processing	Glazing (surface finishing)	Reduced surface roughness gap vs. milled	Not linked to a specific failure mode in the synthesis text	n = 1; Hassan et al., 2025 [31]
Material/Interface	Resin cement bonding to zirconia (surface conditioning + cementation protocols)	Not specified	Adhesive failure predominant in bond-strength testing	n = 3; Cho et al., 2025 [43]; Bömicke et al., 2024 [59]; Zhao et al., 2025 [53]
Failure context	Mechanical testing context (crown fracture/cyclic loading studies)	Not specified	Catastrophic fracture dominant failure mode (printed and milled)	n = 3; Refaie et al., 2023a [30]; Ioannidis et al., 2020 [37]; Zenthöfer et al., 2024b [58]
Failure context	Implant-supported crown configuration	Not specified	Fracture at the abutment–implant interface	n = 1; Zandinejad et al., 2021 [20]
Failure context	Tooth-supported restoration–cement–tooth complex	Not specified	Cohesive fracture within the restoration–cement–tooth complex	n = 1; Ioannidis et al., 2020 [37]

Abbreviations: AM, additive manufacturing; NPJ, nanoparticle jetting; DLP, digital light processing; FDM, fused deposition modeling.

).

Where etiologic explanations extended beyond direct observation (e.g., attributing fracture behavior to specific processing-induced defects), these interpretations were treated as hypothesized unless directly evidenced by the reported analyses.

4. Discussion

This scoping review mapped a zirconia-dominant and largely preclinical evidence base for directly additively manufactured dental ceramics used in single-unit restorations. Most included studies were laboratory investigations, and most evidence clustered in the R1 and R2 readiness tiers, indicating that the field has progressed beyond isolated material screening but remains concentrated at the level of restoration-relevant in vitro evaluation rather than mature clinical translation. Crowns were the most extensively studied indication, whereas veneers and especially partial-coverage restorations were supported by a smaller and less clinically advanced body of evidence. Although several studies benchmarked AM zirconia against CAD/CAM-milled comparators for fit, accuracy, and mechanical performance, no R5 evidence was identified, underscoring that the current literature supports technical feasibility more strongly than routine clinical adoption [26,36,37].

By indication, crowns form the largest cluster and include the highest-readiness evidence [29,33]. Crown studies most often reported marginal/internal adaptation and manufacturing accuracy (trueness/precision), frequently alongside fracture resistance or fatigue outcomes and commonly benchmarked against CAD/CAM-milled zirconia [30,31,45,60]. Fit/adaptation was assessed with diverse methods (e.g., silicone replica, digital superimposition, micro-CT), limiting direct comparison of absolute values across studies and AM systems [32,35,40]. Aging protocols varied from no aging to thermocycling and extended mechanical cycling, which further constrains synthesis of durability [31,47]. The highest readiness tier for crowns was R4, supported by a randomized trial of nanoparticle jetting zirconia crowns and a self-controlled clinical trial of gel-deposited zirconia crowns, with an additional short-term feasibility pilot using selective laser melting [26,27,29]. Therefore, the current crown literature suggests increasing technical maturity, but not yet broad clinical equivalence across AM routes.

For veneers, the mapped evidence was smaller and was dominated by in vitro laminate/ultrathin veneer studies, with one retrospective comparative clinical study providing the highest-readiness evidence (R4) [28,52,53,55]. Veneer outcomes centered on marginal adaptation/fit, fatigue or fracture resistance under simulated aging, and surface- or aging-related vulnerability relevant to thin ceramic sections [52,53,55,61]. The clinical veneer evidence compared gel-deposited self-glazed zirconia veneers with lithium disilicate veneers over multi-year follow-up, offering the most clinically proximal information in this indication [28].

For partial-coverage restorations, evidence was concentrated in occlusal tabletop and partial-crown scenarios evaluated in vitro (R2) [34,37,38,44]. These studies emphasized marginal/internal adaptation and load-bearing behavior under static or fatigue loading, often contextualized against milled zirconia and/or heat-pressed lithium disilicate [34,37,38]. No clinical feasibility or comparative trials were identified for additively manufactured zirconia partial-coverage restorations in the mapped evidence, leaving a translational gap for minimally invasive indications.

From a clinical perspective, the mapped evidence suggests that directly additively manufactured zirconia restorations should still be interpreted relative to the established CAD/CAM benchmark rather than as a fully equivalent replacement workflow. Existing studies indicate that AM zirconia crowns can achieve restoration-level outcomes that are increasingly relevant to practice, particularly for fit/adaptation, manufacturing accuracy, and fracture-related testing. However, the evidence remains heterogeneous with respect to printers, materials, build orientation, post-processing, finishing, and aging design. For clinicians and laboratories, this means that geometry alone is not an adequate basis for judging clinical readiness; reproducibility, reliability, and reporting transparency remain central to translation.

Route-specific patterns help explain where the evidence accumulates and where it dwindles. Nanoparticle jetting (NPJ) contributed both clinical crown data and restoration-level in vitro work focusing on accuracy and margin quality [26,33,35]. Vat photopolymerization of zirconia slurries (LCM/DLP/SLA) accounted for the largest share of restoration-level studies across indications, with repeated attention to fit, trueness, and fracture behavior relative to milling [36,45,57,60]. In this route, several studies directly tested build direction and/or system effects on adaptation and accuracy, highlighting orientation as a recurrent determinant [43,57]. Furthermore, slurry-based ceramic printing introduces a densification step with substantial shrinkage, and defect control during debinding/sintering has been highlighted as a central constraint on dimensional fidelity and strength in zirconia AM [13,14]. Consistent with this, formulation and solids-loading effects have been shown to influence microstructure and physical properties after additive processing and sintering [14,62]. Gel/deposition-based approaches provide the available R4 evidence for both crowns and veneers and therefore represent an important clinical translation pathway in the current map [27,28,50]. Other AM approaches (e.g., direct ink writing/robocasting and filament-based methods) were mainly represented at foundational tiers, contributing coupon/disk mechanical and microstructural characterization rather than restoration-level validation [41,42,56].

The evidence map also highlights other underreported factors beyond laboratory performance. Few included studies described workflow efficiency, operator burden, cost implications, or standardization across printing and post-processing stages in sufficient detail to inform real-world implementation. In addition, the rapid evolution of commercial hardware, ceramic formulations, and software ecosystems means that published evidence may lag behind current technical capabilities. This reinforces the need for clearer reporting standards and clinically oriented comparative studies that evaluate not only restoration outcomes, but also the practical conditions under which AM ceramic workflows can be integrated into routine prosthodontic care.

The failure pathway synthesis suggests that reported failures in restoration-level studies most often manifest as fracture under static or cyclic loading, while the underlying defect signatures are not consistently characterized [19,23,30]. Where microstructural and phase analyses were performed, studies reported features consistent with brittle-flaw control, including porosity/density variation, microstructural heterogeneity, and surface topography related to printing and finishing [12,28,40,41,50,53]. Porosity and volumetric change have been examined in printed zirconia, supporting porosity as a possible mediator of both dimensional outcomes and mechanical variability [40]. Orientation-dependent surface roughness and strength differences also support a pathway in which surface flaw populations and anisotropy influence load-bearing behavior, particularly for thin restorations [1,18]. Mean strength by itself may not reflect the likelihood of clinical failure, because two materials with similar averages can differ greatly in variability and reliability; accordingly, Weibull analyses are often used to compare how AM affects both characteristic strength and the spread of strength values [63,64].

When situated within prior literature, the mapped evidence aligns with zirconia-focused syntheses that emphasize fit/trueness, flexural strength, and the influence of orientation, layer thickness, and post-processing variables [1,13,16,17,23]. The additional contribution of the present scoping approach is the explicit linkage of evidence volume to indication and readiness tier, which clarifies that clinical evidence is limited relative to laboratory validation. Within the mapped primary studies, lithium disilicate and other glass-ceramics appeared chiefly as conventionally manufactured comparators rather than as directly printed definitive restorations, consistent with the zirconia-centric focus of contemporary AM dental ceramics reviews [34,37,43].

From a clinical-readiness perspective, the concentration of evidence in R1 and R2 indicates that many AM ceramic workflows have advanced through foundational and restoration-level laboratory validation, but have not yet been tested extensively in clinical settings [32,33,65]. The available R3 and R4 studies provide encouraging feasibility and comparative signals, but follow-up remains short- to medium-term and is insufficient to characterize longer-term complications, consistency across operators, and performance across routine care settings [26,27,28,29]. Progress toward higher readiness will likely depend on better harmonization of metrology endpoints, aging and fatigue protocols, and more complete reporting of processing variables such as orientation, layer thickness, debinding and sintering schedules, shrinkage compensation, and finishing [1,13,14].

Several limitations should be considered when interpreting this evidence map. First, as a scoping review, this study was designed to map the extent, nature, and maturity of the available evidence rather than to generate pooled effect estimates or definitive equivalence claims across AM routes. Second, heterogeneity in fit and trueness metrics, mechanical testing configurations, aging protocols, and reporting practices limited cross-study comparability. Third, no formal critical appraisal was undertaken, so the mapped patterns should be interpreted primarily as indicators of evidence maturity rather than certainty of effect. Finally, despite comprehensive searching, relevant studies may still have been missed because of database coverage limits, English-language restriction, indexing delays, evolving terminology, and the practical limits of supplementary AI-assisted retrieval. Rapid iteration in printers, ceramic formulations, and software may also mean that the published evidence lags behind current commercial capabilities.

A focused research agenda for improving clarity and accelerating translation includes consistent reporting of AM route and printer/system parameters (orientation, layer thickness, support strategy); feedstock/slurry characteristics (solids loading and binder system where available); debinding and sintering schedules and shrinkage compensation; surface finishing and internal conditioning; cementation protocols for bonded restorations; defect characterization (e.g., micro-CT porosity, SEM microcracks, XRD phase); and parameterized aging regimens with clinically meaningful endpoints in clinical studies [13,53,59]. Furthermore, the reporting of reliability metrics alongside mean outcomes (e.g., Weibull modulus with flexural strength) may better reflect readiness for thin restorations and minimally invasive indications [63,66].

Directly additively manufactured zirconia restorations now have expanding restoration-level in vitro evidence base across NPJ, vat photopolymerization slurry printing, and gel/deposition routes, but clinical evidence remains limited and short-term. Crowns currently represent the most advanced indication in readiness terms, whereas veneers and partial-coverage restorations remain supported by a smaller and less clinically mature body of evidence. Taken together, the mapped literature suggests progressive technical development, but not yet broad clinical maturity.

5. Conclusions

In summary, the available literature on directly additively manufactured dental ceramics for crowns, veneers, and partial-coverage restorations is concentrated in zirconia-based, preclinical research, with crowns representing the most mature indication. Across the 35 included sources, most evidence clustered in the R1 and R2 tiers, only a small number of studies reached R3 or R4, and no R5-level evidence was identified. The mapped evidence supports technical feasibility and increasing restoration-level evaluation, but clinical translation remains limited by sparse clinical datasets, heterogeneous testing and reporting practices, and incomplete characterization of process-linked defect pathways. At present, the field is better described as progressing toward clinical readiness than as having achieved broad clinical maturity. Future studies should prioritize standardized reporting of AM and post-processing parameters, clinically relevant aging and fatigue protocols, and larger, longer-term comparative clinical evaluations.