{"id":"d07fc49a-ff86-47b5-917b-a1f4827a964f","arxiv_id":"2507.02837","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For a one-phase tumor growth free boundary problem with an obstacle, viscosity solutions exist and the free boundary is analytic in the interior and C^{1,α} at contact points under a non-degeneracy condition.","lead":"This paper proves existence and regularity for a tumor growth free boundary model in which the tumor moves along a fixed direction and cannot enter an obstacle. It shows the tumor boundary is analytic away from the obstacle and C^{1,α} smooth at points where it touches the obstacle, when the direction field pushes the tumor toward the obstacle.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 5.8's claimed equality σd=σW is false for non-vertical W; for v=-xd, σd=-1 while σW=-Wd, so the proof of Theorem 1.4 contains a gap that should be corrected before acceptance.","rationale":"The reader's weakest_assumption (non-degeneracy V·ν>0) is an explicit hypothesis of the main theorems, not a hidden flaw, so I do not share it as the primary concern. My stress-test instead focused on the internal correctness of the regularity theory for the oblique thin obstacle problem (Theorem 1.4), which is the engine behind Theorem 1.3. In Section 5.3, Lemma 5.8 asserts a pointwise equality between the normal derivative and the oblique derivative along W on the contact set. This assertion is not consistent with the geometry of oblique or even non-unit directions: a simple, legitimate solution -xd gives σd=-1 and σW=-Wd. The proof of the lemma itself only establishes that the tangential derivative contribution vanishes, yielding σW=Wd σd. The subsequent use in Lemma 5.10, where the vertical expansion is written with σW in place of σd, carries this mistake into the Hölder estimate for σW. Because Wd≥δ>0, the missing factor is bounded away from 0 and infinity, so the estimates are repairable by adjusting constants; nevertheless, as written, the proof of the central regularity theorem contains a false statement. I therefore recommend a CONDITIONAL acceptance: the authors should correct Lemma 5.8 (or replace σd/σW by the appropriate relation) and verify that the constants in Section 5.4 still close, or provide a direct proof of the C^{0,α} regularity of σW without this identity. This is a concrete, localized issue rather than a global collapse of the argument.","tokens_in":56106,"tokens_out":35874,"duration_ms":363990,"concrete_test":"Check the identity in Lemma 5.8 against the explicit viscosity solution v(x)=-xd of (5.1) with W=(W',Wd), Wd∈(0,1). For this solution, σd = -1 and σW = -Wd, so σd=σW only if Wd=1. Then re-derive Lemma 5.10 with the corrected relation σd=σW/Wd: verify that the inequalities in the two steps still produce the claimed ball B'_{C1γ}(x)∩G⊂Sγ (with constants adjusted by 1/δ), and that Lemma 5.12's induction still closes. If the corrected proof closes, the main theorem is sound after a minor revision; if not, Theorem 1.4 is in jeopardy.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"In Section 5.3, Lemma 5.8 claims that for a.e. point on the contact set Λ(vℓ), the normal derivative σd(x') := lim_{t→0+} ∂_d vℓ(x'+t eW) equals the oblique derivative σW(x') := lim_{t→0+} ∇vℓ(x'+t eW)·W, where eW=(-W',Wd). This is not an identity: since W = (W',Wd), ∇v·W = W'·∇'v + Wd ∂_d v. The proof of Lemma 5.8 only shows that the tangential contribution W'·∇'v vanishes in the limit, giving σW = Wd σd, hence σd = σW/Wd, not σd = σW. The distinction is not merely notational: the function v(x) = -xd is a viscosity solution of the oblique thin obstacle problem (5.1) (it is harmonic in B+1, v=0≥0 on B'1, and ∇v·W = -Wd ≤ 0; Ω_v is empty). For this solution, σd = -1 but σW = -Wd, so equality holds only in the normal case Wd=1. Lemma 5.10 and the C^{0,α} estimate for σW in Lemma 5.12 rely on the displayed expansion vℓ(z',yd)=vℓ(z',0)+σd(z')yd+... with σd replaced by σW; if the correct factor 1/Wd is inserted, the constants change by at most 1/δ but the arguments still go through. As written, however, the proof of Theorem 1.4—the key input to Theorem 1.3—contains a false assertion.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a one-phase free boundary problem modelling tumor growth in a domain with an obstacle. The tumor region is {u>0}, u is harmonic in its interior, and the free boundary satisfies |∇u|^2 = ∇u·V away from the obstacle, with an inequality on the obstacle. The authors prove existence of viscosity solutions by Perron's method, interior analyticity of the free boundary near nondegenerate regular points, and C^{1,α} regularity of the free boundary at contact points with the obstacle, assuming a nondegeneracy condition V·ν_K>0. The contact-point result is reduced to a new C^{1,α} regularity theorem for an oblique thin obstacle problem, which is proved in Section 5.","tokens_in":56445,"tokens_out":9082,"duration_ms":110690,"significance":"If the main theorems are correct, this is a substantial contribution to the regularity theory of nonvariational free boundary problems arising from tumor growth. The architecture is coherent and mostly self-contained: Perron existence, improvement-of-flatness arguments, a hodograph transform for analyticity, and a reduction of the contact case to an oblique thin obstacle problem. The paper is honest about its main structural limitation, namely that the nondegeneracy assumptions V·ν>0 and V·ν_K(x0)>0 are essential and the degenerate case is not treated. The central proofs are detailed and the claims are concrete and checkable. However, the proof of Theorem 1.4, which is the key input for Theorem 1.3, contains a false identity in Lemma 5.8; the error appears fixable, but the manuscript cannot be accepted in its present form.","major_comments":[{"comment":"The asserted equality σd = σW is false for general oblique W. Writing W=(W',Wd) and eW=(-W',Wd), one has ∇v·W = Wd ∂d v + W'·∇'v. The proof of Lemma 5.8 shows only that the tangential contribution W'·∇'v tends to zero at almost every contact point, so the correct conclusion is σW = Wd σd, not σd = σW. The counterexample v(x)=-xd is a viscosity solution of (5.1): it is harmonic, v=0≥0 on B'1, ∇v·W = -Wd ≤ 0, and Ω_v is empty; for this solution σd = -1 while σW = -Wd, so equality holds only in the normal case Wd=1. The identity is used in Lemma 5.10 and in the C^{0,α} argument for σW in Lemma 5.12 through the expansion vℓ(z',yd) = vℓ(z',0)+σW(z')yd+... . Inserting the correct factor 1/Wd changes constants by at most a factor depending on 1/δ, and Wd≥δ, so the argument can likely be repaired; nevertheless, as written the proof of Theorem 1.4 contains a false assertion and must be corrected.","section":"Section 5.3, Lemma 5.8"},{"comment":"The penalized problems v_k^ℓ are introduced and the properties (a)-(e) are asserted 'exactly as in [51, Section 2]', but the existence of the penalized solutions and, more importantly, the uniform bound (b) on ∇v_k^ℓ·W are not proved in the manuscript. Since the oblique condition has a nontrivial tangential component, the transfer of the penalization argument from [51] is not immediate and needs a precise justification or a statement-by-statement reference. This is a load-bearing step in the proof that σW≤0, which is used in Lemma 5.12 and in the proof of Theorem 5.1.","section":"Section 5.3, Lemma 5.7"}],"minor_comments":[{"comment":"The set G is introduced only inside the conclusion of the lemma; please state explicitly that there exists a Borel set G⊂B'1 with H^{d-1}(B'1\\G)=0 such that the displayed identity holds on Λ(vℓ)∩G.","section":"Section 5.3, Lemma 5.8"},{"comment":"There are several unpolished formulations, e.g. 'Differently form the interior version' in Section 1.4, 'a non vanishing function' in Section 3.1.1, and 'Providing that' in Section 5.4; these should be corrected in a final revision.","section":"Section 1.4 and Section 5.4"},{"comment":"The assumptions V·ν>0 and V·ν_K(x0)>0 are essential for the proofs, and the borderline case V·ν=0 is explicitly outside the scope; it would help future readers if the paper stated in one place that the degenerate case remains open.","section":"Section 1.3.2 and Theorems 1.2-1.3"}],"recommendation":"major_revision","confidential_remarks":"The false identity in Lemma 5.8 appears to be an algebraic slip rather than a conceptual flaw; with the factor 1/Wd inserted, the estimates should still close because Wd≥δ. I do not recommend rejection. Please also ask the authors to supply the missing details for Lemma 5.7, since the current proof delegates a central penalization bound to [51] without a precise transfer statement."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The main results are new and the paper is well built around them. Theorem 1.3, C^{1,α} regularity at obstacle contact points, depends on a genuinely new auxiliary result, Theorem 1.4 for the oblique thin obstacle problem. The authors correctly note that the nonlocal result in [32] does not transfer, and their proof via semiconvexity, the adjoint direction, and a C^{0,α} estimate for the oblique derivative is a real contribution worth having on record. The existence theory via Perron's method, the interior flatness improvement, and the hodograph step to analyticity are all executed cleanly, and the paper is honest about the non-degeneracy assumptions V·ν > 0, stating them explicitly and never pretending to cover the degenerate case. Citation patterns look fair; the self-citations are to closely related work and are used contextually, not as a crutch. No code or formal verification, but this is a PDE paper where the proofs are the deliverable, so that is not a gap.\n\nThere is one genuine soft spot: Lemma 5.8 is false as stated for non-vertical W. The claimed identity σd = σW on the contact set should be σW = Wd σd, because the tangential contribution vanishes only in the limit and the normal component survives. The test solution v(x) = -xd makes this concrete: σd = -1 but σW = -Wd. As far as I can tell the downstream arguments in Lemma 5.10 and Lemma 5.12 use the wrong identity in the displayed expansions, but the correction only introduces a factor 1/Wd, bounded by 1/δ. So the architecture of the proof survives, but the current text contains a false assertion in a load-bearing lemma, and the authors need to rewrite that section before the paper is publishable. The other delegated claims, to [51] and [58], are standard enough but not fully self-contained, which I'd rank as minor.\n\nThis is a solid paper for people working in free boundary regularity: it opens a new model with an obstacle and a preferred direction, and the oblique thin obstacle theorem is a reusable tool. I would send it to a serious referee with a request to check the corrected Section 5, not desk-reject it. If the authors fix Lemma 5.8, I would accept it.","headline":"A genuinely new and mostly rigorous free boundary paper; the oblique thin obstacle result is real and useful, but Lemma 5.8 has a fixable false equality that the authors need to correct.","tokens_in":56987,"tokens_out":1527,"would_cite":true,"duration_ms":20847,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35R35","35Q92","76D27"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that the free boundary of a tumor growth model is well behaved: analytic away from obstacles and C^1,α where it meets them.","keywords":["free boundary problem","viscosity solutions","Perron's method","thin obstacle problem","oblique boundary condition","C^{1,\\alpha} regularity","tumor growth","Hele-Shaw problem"],"falsifier":"Exhibit a viscosity solution with a contact point where the growth direction is exactly tangent to the obstacle (V·ν_K = 0) whose free boundary is not a single $C^{1}$,α graph—for instance, a boundary that splits into two branches or develops a tangential cusp. Within the oblique thin obstacle problem, the corresponding test is whether the $C^{1}$,α estimate fails as the obliqueness W·e_d approaches zero, or whether a blow-up at a point with W·e_d = 0 yields a non-$C^{1}$ limit.","tokens_in":55894,"feed_emoji":"🧫","tokens_out":4042,"duration_ms":42739,"temperature":0.7,"pith_summary":"This paper develops an existence and regularity theory for a geometric free boundary problem describing a tumor that grows along a preferred direction inside an accessible region while being blocked by an obstacle. The central claim is that the free boundary, the tumor surface, is well behaved everywhere it can appear: analytic away from the obstacle, and $C^{1}$,α where it meets the obstacle. The proof works through Perron's method for existence, an improvement-of-flatness argument for interior points, and a reduction of contact-point regularity to a new $C^{1}$,α estimate for an oblique thin obstacle problem. A sympathetic reader should care because the result turns a non-variational, drift-driven free boundary condition into classical regularity theory.","feed_headline":"Tumor-model boundary is smooth where it hits an obstacle","feed_subtitle":"New proof shows C^1,α regularity at obstacle contacts and analyticity away from them.","key_machinery":"The load-bearing object is the oblique thin obstacle problem: Δv = 0 in the half ball, with oblique Neumann condition ∇v·W = 0 off the contact set and ∇v·W ≤ 0 on the thin space, where W·e_d ≥ δ > 0. Its $C^{1}$,α regularity (Theorem 1.4) is the engine; the contact-point proof flattens the obstacle by a change of coordinates, shows every contact point has a planar blow-up, and then feeds an improvement-of-flatness iteration at branching points whose limiting linearized problem is exactly this oblique thin obstacle problem.","core_discovery":"The paper establishes that viscosity solutions of the obstacle-tumor free boundary problem exist for continuous boundary data and that their free boundary has the following structure: at interior regular points where the growth direction has positive normal component, the boundary is analytic; at contact points with the obstacle where the growth direction points into the obstacle, the boundary is a $C^{1}$,α graph for every α below a universal exponent. The key reduction is Theorem 1.3: the boundary behavior at contact points is controlled by the regularity of the oblique thin obstacle problem, for which the paper proves $C^{1}$,α estimates with a constant depending only on the obliqueness.","pith_inferences":["The paper leaves the degenerate case where the growth direction is exactly tangent to the obstacle (V·ν_K = 0) open; at such points the improvement-of-flatness argument loses its uniform control and genuinely different singular behavior might occur, such as cusps aligned with the drift.","The authors expect the optimal exponent in the contact theorem to coincide with the optimal regularity of the oblique thin obstacle problem; if the oblique problem is shown to be C^1,1/2-sharp as in the classical thin obstacle case, the contact theorem would inherit that sharpness.","A rigorous derivation of the model as an incompressible limit of a porous-medium system is conjectured but not proved; supplying that limit would connect this static regularity theory to the time-dependent Hele-Shaw flow.","A numerical testable extension: simulate the model with V nearly tangent to the obstacle and check whether the contact boundary loses Hölder regularity as V·ν_K approaches zero."],"forward_implications":["If the contact regularity theorem holds, the tumor boundary cannot develop cusps or fractal singularities at obstacle contacts, so the interface can be reliably resolved numerically.","The reduction to the oblique thin obstacle problem means any future sharpening of the exponent in Theorem 1.4 immediately upgrades the contact regularity theorem.","The interior result gives analyticity of the free boundary away from the obstacle, so the interface is not merely differentiable but fully smooth there.","Existence via Perron's method provides a canonical viscosity solution for arbitrary continuous boundary data, giving a platform for uniqueness or comparison questions.","Combined with the graphical property, the free boundary is a graph along the drift direction, which constrains the possible shapes of the tumor."],"supporting_citations":[{"why":"Supplies the barrier, semiconvexity, and σ-function strategy for the classical thin obstacle problem that Section 5 adapts to the oblique case.","marker":"[10]"},{"why":"Provides Perron's method for existence and the contact-point expansion lemma used in Proposition 4.5.","marker":"[12]"},{"why":"Provides the boundary improvement-of-flatness framework for one-phase free boundaries that Section 4.3 adapts to contact points.","marker":"[17]"},{"why":"Supplies the interior improvement-of-flatness and partial Harnack inequality used in Section 3.","marker":"[26]"},{"why":"Gives the oblique-derivative Schauder estimates used for the linearized problem and for the final bootstrap to C^1,α.","marker":"[49]"},{"why":"Provides the penalization argument and the C^1,α scheme for the nonlinear Signorini problem adapted in Lemmas 5.7 and 5.11.","marker":"[51]"},{"why":"Supplies the geometric projection argument used in Proposition 4.17 to conclude the free boundary is a C^1,α graph.","marker":"[59]"}],"fun_headline_variants":["Tumor free boundary is C^1,α at obstacle contact","Tumor model: free boundary meets obstacle as C^1,α graph","Obstacle contact: tumor boundary proven C^1,α","Proof: tumor boundary smooth at obstacle, C^1,α","Tumor growth: C^1,α regularity at obstacle contact"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything rests on the standing non-degeneracy that the growth direction has a uniformly positive normal component on the relevant part of the free boundary, in particular at obstacle contacts, so that flatness cannot degenerate to zero slope; the theory says nothing at points where the direction is exactly tangent or the slope vanishes.","fun_headline_variants_meta":{"raw":{"variants":["Tumor free boundary is C^1,α at obstacle contact","Tumor model: free boundary meets obstacle as C^1,α graph","Obstacle contact: tumor boundary proven C^1,α","Proof: tumor boundary smooth at obstacle, C^1,α","Tumor growth: C^1,α regularity at obstacle contact"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000529,"raw_usage":{"total_tokens":2482,"prompt_tokens":809,"completion_tokens":1673,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":425,"completion_tokens_details":{"reasoning_tokens":1580}},"tokens_in":425,"tokens_out":1673,"duration_ms":13817,"temperature":1.0,"reasoning_tokens":1580,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T20:22:20.861484+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a viscosity solution with a contact point where the growth direction is exactly tangent to the obstacle (V·ν_K = 0) whose free boundary is not a single $C^{1}$,α graph—for instance, a boundary that splits into two branches or develops a tangential cusp. Within the oblique thin obstacle problem, the corresponding test is whether the $C^{1}$,α estimate fails as the obliqueness W·e_d approaches zero, or whether a blow-up at a point with W·e_d = 0 yields a non-$C^{1}$ limit.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the barrier, semiconvexity, and σ-function strategy for the classical thin obstacle problem that Section 5 adapts to the oblique case."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides Perron's method for existence and the contact-point expansion lemma used in Proposition 4.5."},{"cited_title":"Boundary regularity for the free boundary in the one-phase problem","cited_arxiv_id":null,"evidence_quote":"Provides the boundary improvement-of-flatness framework for one-phase free boundaries that Section 4.3 adapts to contact points."},{"cited_title":"De Silva","cited_arxiv_id":null,"evidence_quote":"Supplies the interior improvement-of-flatness and partial Harnack inequality used in Section 3."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the oblique-derivative Schauder estimates used for the linearized problem and for the final bootstrap to C^1,α."},{"cited_title":"Milakis and L","cited_arxiv_id":null,"evidence_quote":"Provides the penalization argument and the C^1,α scheme for the nonlinear Signorini problem adapted in Lemmas 5.7 and 5.11."},{"cited_title":"Velichkov","cited_arxiv_id":null,"evidence_quote":"Supplies the geometric projection argument used in Proposition 4.17 to conclude the free boundary is a C^1,α graph."}],"review_version":1}